basic.py 186 KB
Newer Older
wxchan's avatar
wxchan committed
1
# coding: utf-8
2
"""Wrapper for C API of LightGBM."""
3
import abc
wxchan's avatar
wxchan committed
4
import ctypes
5
import inspect
6
import json
wxchan's avatar
wxchan committed
7
import warnings
8
from collections import OrderedDict
9
from copy import deepcopy
10
from enum import Enum
11
from functools import wraps
12
from os import SEEK_END, environ
13
14
from os.path import getsize
from pathlib import Path
15
from tempfile import NamedTemporaryFile
16
from typing import TYPE_CHECKING, Any, Callable, Dict, Iterable, List, Optional, Set, Tuple, Union
wxchan's avatar
wxchan committed
17
18
19
20

import numpy as np
import scipy.sparse

21
from .compat import PANDAS_INSTALLED, concat, dt_DataTable, pd_CategoricalDtype, pd_DataFrame, pd_Series
wxchan's avatar
wxchan committed
22
23
from .libpath import find_lib_path

24
25
26
if TYPE_CHECKING:
    from typing import Literal

27
28
29
30
31
32
33
34
35
__all__ = [
    'Booster',
    'Dataset',
    'LGBMDeprecationWarning',
    'LightGBMError',
    'register_logger',
    'Sequence',
]

36
_BoosterHandle = ctypes.c_void_p
37
_DatasetHandle = ctypes.c_void_p
38
39
40
41
_ctypes_int_ptr = Union[
    "ctypes._Pointer[ctypes.c_int32]",
    "ctypes._Pointer[ctypes.c_int64]"
]
42
43
44
45
_ctypes_int_array = Union[
    "ctypes.Array[ctypes._Pointer[ctypes.c_int32]]",
    "ctypes.Array[ctypes._Pointer[ctypes.c_int64]]"
]
46
47
48
49
50
51
52
53
_ctypes_float_ptr = Union[
    "ctypes._Pointer[ctypes.c_float]",
    "ctypes._Pointer[ctypes.c_double]"
]
_ctypes_float_array = Union[
    "ctypes.Array[ctypes._Pointer[ctypes.c_float]]",
    "ctypes.Array[ctypes._Pointer[ctypes.c_double]]"
]
54
_LGBM_EvalFunctionResultType = Tuple[str, float, bool]
55
_LGBM_BoosterBestScoreType = Dict[str, Dict[str, float]]
56
_LGBM_BoosterEvalMethodResultType = Tuple[str, str, float, bool]
57
58
_LGBM_CategoricalFeatureConfiguration = Union[List[str], List[int], "Literal['auto']"]
_LGBM_FeatureNameConfiguration = Union[List[str], "Literal['auto']"]
59
60
61
62
63
64
_LGBM_GroupType = Union[
    List[float],
    List[int],
    np.ndarray,
    pd_Series
]
65
66
67
68
_LGBM_PositionType = Union[
    np.ndarray,
    pd_Series
]
69
70
71
72
73
74
75
_LGBM_InitScoreType = Union[
    List[float],
    List[List[float]],
    np.ndarray,
    pd_Series,
    pd_DataFrame,
]
76
77
78
79
80
81
82
83
84
85
86
_LGBM_TrainDataType = Union[
    str,
    Path,
    np.ndarray,
    pd_DataFrame,
    dt_DataTable,
    scipy.sparse.spmatrix,
    "Sequence",
    List["Sequence"],
    List[np.ndarray]
]
87
_LGBM_LabelType = Union[
88
89
    List[float],
    List[int],
90
91
92
93
    np.ndarray,
    pd_Series,
    pd_DataFrame
]
94
95
96
97
98
99
100
101
_LGBM_PredictDataType = Union[
    str,
    Path,
    np.ndarray,
    pd_DataFrame,
    dt_DataTable,
    scipy.sparse.spmatrix
]
102
103
104
105
106
107
_LGBM_WeightType = Union[
    List[float],
    List[int],
    np.ndarray,
    pd_Series
]
108
109
110
ZERO_THRESHOLD = 1e-35


111
112
113
114
def _is_zero(x: float) -> bool:
    return -ZERO_THRESHOLD <= x <= ZERO_THRESHOLD


115
def _get_sample_count(total_nrow: int, params: str) -> int:
116
117
118
    sample_cnt = ctypes.c_int(0)
    _safe_call(_LIB.LGBM_GetSampleCount(
        ctypes.c_int32(total_nrow),
119
        _c_str(params),
120
121
122
123
        ctypes.byref(sample_cnt),
    ))
    return sample_cnt.value

wxchan's avatar
wxchan committed
124

125
126
127
128
129
130
class _MissingType(Enum):
    NONE = 'None'
    NAN = 'NaN'
    ZERO = 'Zero'


131
class _DummyLogger:
132
    def info(self, msg: str) -> None:
133
        print(msg)  # noqa: T201
134

135
    def warning(self, msg: str) -> None:
136
137
138
        warnings.warn(msg, stacklevel=3)


139
140
141
_LOGGER: Any = _DummyLogger()
_INFO_METHOD_NAME = "info"
_WARNING_METHOD_NAME = "warning"
142
143


144
145
146
147
def _has_method(logger: Any, method_name: str) -> bool:
    return callable(getattr(logger, method_name, None))


148
149
150
def register_logger(
    logger: Any, info_method_name: str = "info", warning_method_name: str = "warning"
) -> None:
151
152
153
154
    """Register custom logger.

    Parameters
    ----------
155
    logger : Any
156
        Custom logger.
157
158
159
160
    info_method_name : str, optional (default="info")
        Method used to log info messages.
    warning_method_name : str, optional (default="warning")
        Method used to log warning messages.
161
    """
162
163
164
165
166
167
    if not _has_method(logger, info_method_name) or not _has_method(logger, warning_method_name):
        raise TypeError(
            f"Logger must provide '{info_method_name}' and '{warning_method_name}' method"
        )

    global _LOGGER, _INFO_METHOD_NAME, _WARNING_METHOD_NAME
168
    _LOGGER = logger
169
170
    _INFO_METHOD_NAME = info_method_name
    _WARNING_METHOD_NAME = warning_method_name
171
172


173
def _normalize_native_string(func: Callable[[str], None]) -> Callable[[str], None]:
174
    """Join log messages from native library which come by chunks."""
175
    msg_normalized: List[str] = []
176
177

    @wraps(func)
178
    def wrapper(msg: str) -> None:
179
180
181
182
183
184
185
186
187
188
189
        nonlocal msg_normalized
        if msg.strip() == '':
            msg = ''.join(msg_normalized)
            msg_normalized = []
            return func(msg)
        else:
            msg_normalized.append(msg)

    return wrapper


190
def _log_info(msg: str) -> None:
191
    getattr(_LOGGER, _INFO_METHOD_NAME)(msg)
192
193


194
def _log_warning(msg: str) -> None:
195
    getattr(_LOGGER, _WARNING_METHOD_NAME)(msg)
196
197
198


@_normalize_native_string
199
def _log_native(msg: str) -> None:
200
    getattr(_LOGGER, _INFO_METHOD_NAME)(msg)
201
202


203
def _log_callback(msg: bytes) -> None:
204
    """Redirect logs from native library into Python."""
205
    _log_native(str(msg.decode('utf-8')))
206
207


208
def _load_lib() -> ctypes.CDLL:
209
    """Load LightGBM library."""
wxchan's avatar
wxchan committed
210
211
212
    lib_path = find_lib_path()
    lib = ctypes.cdll.LoadLibrary(lib_path[0])
    lib.LGBM_GetLastError.restype = ctypes.c_char_p
213
    callback = ctypes.CFUNCTYPE(None, ctypes.c_char_p)
214
    lib.callback = callback(_log_callback)  # type: ignore[attr-defined]
215
    if lib.LGBM_RegisterLogCallback(lib.callback) != 0:
216
        raise LightGBMError(lib.LGBM_GetLastError().decode('utf-8'))
wxchan's avatar
wxchan committed
217
218
    return lib

wxchan's avatar
wxchan committed
219

220
221
222
223
224
225
226
# we don't need lib_lightgbm while building docs
_LIB: ctypes.CDLL
if environ.get('LIGHTGBM_BUILD_DOC', False):
    from unittest.mock import Mock  # isort: skip
    _LIB = Mock(ctypes.CDLL)  # type: ignore
else:
    _LIB = _load_lib()
wxchan's avatar
wxchan committed
227

wxchan's avatar
wxchan committed
228

229
_NUMERIC_TYPES = (int, float, bool)
230
_ArrayLike = Union[List, np.ndarray, pd_Series]
231
232


233
def _safe_call(ret: int) -> None:
234
235
    """Check the return value from C API call.

wxchan's avatar
wxchan committed
236
237
238
    Parameters
    ----------
    ret : int
239
        The return value from C API calls.
wxchan's avatar
wxchan committed
240
241
    """
    if ret != 0:
242
        raise LightGBMError(_LIB.LGBM_GetLastError().decode('utf-8'))
wxchan's avatar
wxchan committed
243

wxchan's avatar
wxchan committed
244

245
def _is_numeric(obj: Any) -> bool:
246
    """Check whether object is a number or not, include numpy number, etc."""
wxchan's avatar
wxchan committed
247
248
249
    try:
        float(obj)
        return True
wxchan's avatar
wxchan committed
250
251
252
    except (TypeError, ValueError):
        # TypeError: obj is not a string or a number
        # ValueError: invalid literal
wxchan's avatar
wxchan committed
253
254
        return False

wxchan's avatar
wxchan committed
255

256
def _is_numpy_1d_array(data: Any) -> bool:
257
    """Check whether data is a numpy 1-D array."""
258
    return isinstance(data, np.ndarray) and len(data.shape) == 1
wxchan's avatar
wxchan committed
259

wxchan's avatar
wxchan committed
260

261
def _is_numpy_column_array(data: Any) -> bool:
262
263
264
265
266
267
268
    """Check whether data is a column numpy array."""
    if not isinstance(data, np.ndarray):
        return False
    shape = data.shape
    return len(shape) == 2 and shape[1] == 1


269
def _cast_numpy_array_to_dtype(array: np.ndarray, dtype: "np.typing.DTypeLike") -> np.ndarray:
270
    """Cast numpy array to given dtype."""
271
272
273
274
275
    if array.dtype == dtype:
        return array
    return array.astype(dtype=dtype, copy=False)


276
def _is_1d_list(data: Any) -> bool:
277
    """Check whether data is a 1-D list."""
278
    return isinstance(data, list) and (not data or _is_numeric(data[0]))
wxchan's avatar
wxchan committed
279

wxchan's avatar
wxchan committed
280

281
282
283
def _is_1d_collection(data: Any) -> bool:
    """Check whether data is a 1-D collection."""
    return (
284
        _is_numpy_1d_array(data)
285
        or _is_numpy_column_array(data)
286
        or _is_1d_list(data)
287
288
289
290
        or isinstance(data, pd_Series)
    )


291
292
def _list_to_1d_numpy(
    data: Any,
293
294
    dtype: "np.typing.DTypeLike",
    name: str
295
) -> np.ndarray:
296
    """Convert data to numpy 1-D array."""
297
    if _is_numpy_1d_array(data):
298
        return _cast_numpy_array_to_dtype(data, dtype)
299
    elif _is_numpy_column_array(data):
300
301
        _log_warning('Converting column-vector to 1d array')
        array = data.ravel()
302
        return _cast_numpy_array_to_dtype(array, dtype)
303
    elif _is_1d_list(data):
wxchan's avatar
wxchan committed
304
        return np.array(data, dtype=dtype, copy=False)
305
    elif isinstance(data, pd_Series):
306
        _check_for_bad_pandas_dtypes(data.to_frame().dtypes)
307
        return np.array(data, dtype=dtype, copy=False)  # SparseArray should be supported as well
wxchan's avatar
wxchan committed
308
    else:
309
310
        raise TypeError(f"Wrong type({type(data).__name__}) for {name}.\n"
                        "It should be list, numpy 1-D array or pandas Series")
wxchan's avatar
wxchan committed
311

wxchan's avatar
wxchan committed
312

313
314
315
316
317
318
319
def _is_numpy_2d_array(data: Any) -> bool:
    """Check whether data is a numpy 2-D array."""
    return isinstance(data, np.ndarray) and len(data.shape) == 2 and data.shape[1] > 1


def _is_2d_list(data: Any) -> bool:
    """Check whether data is a 2-D list."""
320
    return isinstance(data, list) and len(data) > 0 and _is_1d_list(data[0])
321
322
323
324
325
326
327
328
329
330
331


def _is_2d_collection(data: Any) -> bool:
    """Check whether data is a 2-D collection."""
    return (
        _is_numpy_2d_array(data)
        or _is_2d_list(data)
        or isinstance(data, pd_DataFrame)
    )


332
333
def _data_to_2d_numpy(
    data: Any,
334
335
    dtype: "np.typing.DTypeLike",
    name: str
336
) -> np.ndarray:
337
338
    """Convert data to numpy 2-D array."""
    if _is_numpy_2d_array(data):
339
        return _cast_numpy_array_to_dtype(data, dtype)
340
341
342
    if _is_2d_list(data):
        return np.array(data, dtype=dtype)
    if isinstance(data, pd_DataFrame):
343
        _check_for_bad_pandas_dtypes(data.dtypes)
344
        return _cast_numpy_array_to_dtype(data.values, dtype)
345
346
347
348
    raise TypeError(f"Wrong type({type(data).__name__}) for {name}.\n"
                    "It should be list of lists, numpy 2-D array or pandas DataFrame")


349
def _cfloat32_array_to_numpy(cptr: "ctypes._Pointer", length: int) -> np.ndarray:
350
    """Convert a ctypes float pointer array to a numpy array."""
wxchan's avatar
wxchan committed
351
    if isinstance(cptr, ctypes.POINTER(ctypes.c_float)):
352
        return np.ctypeslib.as_array(cptr, shape=(length,)).copy()
wxchan's avatar
wxchan committed
353
    else:
354
        raise RuntimeError('Expected float pointer')
wxchan's avatar
wxchan committed
355

Guolin Ke's avatar
Guolin Ke committed
356

357
def _cfloat64_array_to_numpy(cptr: "ctypes._Pointer", length: int) -> np.ndarray:
358
    """Convert a ctypes double pointer array to a numpy array."""
Guolin Ke's avatar
Guolin Ke committed
359
    if isinstance(cptr, ctypes.POINTER(ctypes.c_double)):
360
        return np.ctypeslib.as_array(cptr, shape=(length,)).copy()
Guolin Ke's avatar
Guolin Ke committed
361
362
363
    else:
        raise RuntimeError('Expected double pointer')

wxchan's avatar
wxchan committed
364

365
def _cint32_array_to_numpy(cptr: "ctypes._Pointer", length: int) -> np.ndarray:
366
    """Convert a ctypes int pointer array to a numpy array."""
wxchan's avatar
wxchan committed
367
    if isinstance(cptr, ctypes.POINTER(ctypes.c_int32)):
368
        return np.ctypeslib.as_array(cptr, shape=(length,)).copy()
wxchan's avatar
wxchan committed
369
    else:
370
371
372
        raise RuntimeError('Expected int32 pointer')


373
def _cint64_array_to_numpy(cptr: "ctypes._Pointer", length: int) -> np.ndarray:
374
375
    """Convert a ctypes int pointer array to a numpy array."""
    if isinstance(cptr, ctypes.POINTER(ctypes.c_int64)):
376
        return np.ctypeslib.as_array(cptr, shape=(length,)).copy()
377
378
    else:
        raise RuntimeError('Expected int64 pointer')
wxchan's avatar
wxchan committed
379

wxchan's avatar
wxchan committed
380

381
def _c_str(string: str) -> ctypes.c_char_p:
382
    """Convert a Python string to C string."""
wxchan's avatar
wxchan committed
383
384
    return ctypes.c_char_p(string.encode('utf-8'))

wxchan's avatar
wxchan committed
385

386
def _c_array(ctype: type, values: List[Any]) -> ctypes.Array:
387
    """Convert a Python array to C array."""
388
    return (ctype * len(values))(*values)  # type: ignore[operator]
wxchan's avatar
wxchan committed
389

wxchan's avatar
wxchan committed
390

391
def _json_default_with_numpy(obj: Any) -> Any:
392
393
394
395
396
397
398
399
400
    """Convert numpy classes to JSON serializable objects."""
    if isinstance(obj, (np.integer, np.floating, np.bool_)):
        return obj.item()
    elif isinstance(obj, np.ndarray):
        return obj.tolist()
    else:
        return obj


401
402
403
404
405
406
407
408
def _to_string(x: Union[int, float, str, List]) -> str:
    if isinstance(x, list):
        val_list = ",".join(str(val) for val in x)
        return f"[{val_list}]"
    else:
        return str(x)


409
def _param_dict_to_str(data: Optional[Dict[str, Any]]) -> str:
410
    """Convert Python dictionary to string, which is passed to C API."""
411
    if data is None or not data:
wxchan's avatar
wxchan committed
412
413
414
        return ""
    pairs = []
    for key, val in data.items():
415
        if isinstance(val, (list, tuple, set)) or _is_numpy_1d_array(val):
416
            pairs.append(f"{key}={','.join(map(_to_string, val))}")
417
        elif isinstance(val, (str, Path, _NUMERIC_TYPES)) or _is_numeric(val):
418
            pairs.append(f"{key}={val}")
419
        elif val is not None:
420
            raise TypeError(f'Unknown type of parameter:{key}, got:{type(val).__name__}')
wxchan's avatar
wxchan committed
421
    return ' '.join(pairs)
422

wxchan's avatar
wxchan committed
423

424
class _TempFile:
425
426
    """Proxy class to workaround errors on Windows."""

427
428
429
    def __enter__(self):
        with NamedTemporaryFile(prefix="lightgbm_tmp_", delete=True) as f:
            self.name = f.name
430
            self.path = Path(self.name)
431
        return self
wxchan's avatar
wxchan committed
432

433
    def __exit__(self, exc_type, exc_val, exc_tb):
434
435
        if self.path.is_file():
            self.path.unlink()
436

wxchan's avatar
wxchan committed
437

438
class LightGBMError(Exception):
439
440
    """Error thrown by LightGBM."""

441
442
443
    pass


444
445
446
447
448
449
450
451
# DeprecationWarning is not shown by default, so let's create our own with higher level
class LGBMDeprecationWarning(UserWarning):
    """Custom deprecation warning."""

    pass


class _ConfigAliases:
452
453
454
455
    # lazy evaluation to allow import without dynamic library, e.g., for docs generation
    aliases = None

    @staticmethod
456
    def _get_all_param_aliases() -> Dict[str, List[str]]:
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
        buffer_len = 1 << 20
        tmp_out_len = ctypes.c_int64(0)
        string_buffer = ctypes.create_string_buffer(buffer_len)
        ptr_string_buffer = ctypes.c_char_p(*[ctypes.addressof(string_buffer)])
        _safe_call(_LIB.LGBM_DumpParamAliases(
            ctypes.c_int64(buffer_len),
            ctypes.byref(tmp_out_len),
            ptr_string_buffer))
        actual_len = tmp_out_len.value
        # if buffer length is not long enough, re-allocate a buffer
        if actual_len > buffer_len:
            string_buffer = ctypes.create_string_buffer(actual_len)
            ptr_string_buffer = ctypes.c_char_p(*[ctypes.addressof(string_buffer)])
            _safe_call(_LIB.LGBM_DumpParamAliases(
                ctypes.c_int64(actual_len),
                ctypes.byref(tmp_out_len),
                ptr_string_buffer))
474
        return json.loads(
475
            string_buffer.value.decode('utf-8'),
476
            object_hook=lambda obj: {k: [k] + v for k, v in obj.items()}
477
        )
478
479

    @classmethod
480
481
482
    def get(cls, *args) -> Set[str]:
        if cls.aliases is None:
            cls.aliases = cls._get_all_param_aliases()
483
484
        ret = set()
        for i in args:
485
            ret.update(cls.get_sorted(i))
486
487
        return ret

488
489
490
491
492
493
    @classmethod
    def get_sorted(cls, name: str) -> List[str]:
        if cls.aliases is None:
            cls.aliases = cls._get_all_param_aliases()
        return cls.aliases.get(name, [name])

494
    @classmethod
495
496
497
    def get_by_alias(cls, *args) -> Set[str]:
        if cls.aliases is None:
            cls.aliases = cls._get_all_param_aliases()
498
499
500
501
        ret = set(args)
        for arg in args:
            for aliases in cls.aliases.values():
                if arg in aliases:
502
                    ret.update(aliases)
503
504
505
                    break
        return ret

506

507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
def _choose_param_value(main_param_name: str, params: Dict[str, Any], default_value: Any) -> Dict[str, Any]:
    """Get a single parameter value, accounting for aliases.

    Parameters
    ----------
    main_param_name : str
        Name of the main parameter to get a value for. One of the keys of ``_ConfigAliases``.
    params : dict
        Dictionary of LightGBM parameters.
    default_value : Any
        Default value to use for the parameter, if none is found in ``params``.

    Returns
    -------
    params : dict
        A ``params`` dict with exactly one value for ``main_param_name``, and all aliases ``main_param_name`` removed.
        If both ``main_param_name`` and one or more aliases for it are found, the value of ``main_param_name`` will be preferred.
    """
    # avoid side effects on passed-in parameters
    params = deepcopy(params)

528
529
    aliases = _ConfigAliases.get_sorted(main_param_name)
    aliases = [a for a in aliases if a != main_param_name]
530
531

    # if main_param_name was provided, keep that value and remove all aliases
532
    if main_param_name in params.keys():
533
534
535
        for param in aliases:
            params.pop(param, None)
        return params
536

537
538
539
540
541
    # if main param name was not found, search for an alias
    for param in aliases:
        if param in params.keys():
            params[main_param_name] = params[param]
            break
542

543
544
545
546
547
548
549
    if main_param_name in params.keys():
        for param in aliases:
            params.pop(param, None)
        return params

    # neither of main_param_name, aliases were found
    params[main_param_name] = default_value
550
551
552
553

    return params


554
_MAX_INT32 = (1 << 31) - 1
555

556
"""Macro definition of data type in C API of LightGBM"""
557
558
559
560
_C_API_DTYPE_FLOAT32 = 0
_C_API_DTYPE_FLOAT64 = 1
_C_API_DTYPE_INT32 = 2
_C_API_DTYPE_INT64 = 3
Guolin Ke's avatar
Guolin Ke committed
561

562
"""Matrix is row major in Python"""
563
_C_API_IS_ROW_MAJOR = 1
wxchan's avatar
wxchan committed
564

565
"""Macro definition of prediction type in C API of LightGBM"""
566
567
568
569
_C_API_PREDICT_NORMAL = 0
_C_API_PREDICT_RAW_SCORE = 1
_C_API_PREDICT_LEAF_INDEX = 2
_C_API_PREDICT_CONTRIB = 3
wxchan's avatar
wxchan committed
570

571
"""Macro definition of sparse matrix type"""
572
573
_C_API_MATRIX_TYPE_CSR = 0
_C_API_MATRIX_TYPE_CSC = 1
574

575
"""Macro definition of feature importance type"""
576
577
_C_API_FEATURE_IMPORTANCE_SPLIT = 0
_C_API_FEATURE_IMPORTANCE_GAIN = 1
578

579
"""Data type of data field"""
580
581
582
583
_FIELD_TYPE_MAPPER = {
    "label": _C_API_DTYPE_FLOAT32,
    "weight": _C_API_DTYPE_FLOAT32,
    "init_score": _C_API_DTYPE_FLOAT64,
584
585
    "group": _C_API_DTYPE_INT32,
    "position": _C_API_DTYPE_INT32
586
}
wxchan's avatar
wxchan committed
587

588
"""String name to int feature importance type mapper"""
589
590
591
592
_FEATURE_IMPORTANCE_TYPE_MAPPER = {
    "split": _C_API_FEATURE_IMPORTANCE_SPLIT,
    "gain": _C_API_FEATURE_IMPORTANCE_GAIN
}
593

wxchan's avatar
wxchan committed
594

595
def _convert_from_sliced_object(data: np.ndarray) -> np.ndarray:
596
    """Fix the memory of multi-dimensional sliced object."""
597
    if isinstance(data, np.ndarray) and isinstance(data.base, np.ndarray):
598
        if not data.flags.c_contiguous:
599
600
            _log_warning("Usage of np.ndarray subset (sliced data) is not recommended "
                         "due to it will double the peak memory cost in LightGBM.")
601
602
603
604
            return np.copy(data)
    return data


605
606
607
def _c_float_array(
    data: np.ndarray
) -> Tuple[_ctypes_float_ptr, int, np.ndarray]:
608
    """Get pointer of float numpy array / list."""
609
    if _is_1d_list(data):
wxchan's avatar
wxchan committed
610
        data = np.array(data, copy=False)
611
    if _is_numpy_1d_array(data):
612
        data = _convert_from_sliced_object(data)
613
        assert data.flags.c_contiguous
614
        ptr_data: _ctypes_float_ptr
wxchan's avatar
wxchan committed
615
616
        if data.dtype == np.float32:
            ptr_data = data.ctypes.data_as(ctypes.POINTER(ctypes.c_float))
617
            type_data = _C_API_DTYPE_FLOAT32
wxchan's avatar
wxchan committed
618
619
        elif data.dtype == np.float64:
            ptr_data = data.ctypes.data_as(ctypes.POINTER(ctypes.c_double))
620
            type_data = _C_API_DTYPE_FLOAT64
wxchan's avatar
wxchan committed
621
        else:
622
            raise TypeError(f"Expected np.float32 or np.float64, met type({data.dtype})")
wxchan's avatar
wxchan committed
623
    else:
624
        raise TypeError(f"Unknown type({type(data).__name__})")
625
    return (ptr_data, type_data, data)  # return `data` to avoid the temporary copy is freed
wxchan's avatar
wxchan committed
626

wxchan's avatar
wxchan committed
627

628
629
630
def _c_int_array(
    data: np.ndarray
) -> Tuple[_ctypes_int_ptr, int, np.ndarray]:
631
    """Get pointer of int numpy array / list."""
632
    if _is_1d_list(data):
wxchan's avatar
wxchan committed
633
        data = np.array(data, copy=False)
634
    if _is_numpy_1d_array(data):
635
        data = _convert_from_sliced_object(data)
636
        assert data.flags.c_contiguous
637
        ptr_data: _ctypes_int_ptr
wxchan's avatar
wxchan committed
638
639
        if data.dtype == np.int32:
            ptr_data = data.ctypes.data_as(ctypes.POINTER(ctypes.c_int32))
640
            type_data = _C_API_DTYPE_INT32
wxchan's avatar
wxchan committed
641
642
        elif data.dtype == np.int64:
            ptr_data = data.ctypes.data_as(ctypes.POINTER(ctypes.c_int64))
643
            type_data = _C_API_DTYPE_INT64
wxchan's avatar
wxchan committed
644
        else:
645
            raise TypeError(f"Expected np.int32 or np.int64, met type({data.dtype})")
wxchan's avatar
wxchan committed
646
    else:
647
        raise TypeError(f"Unknown type({type(data).__name__})")
648
    return (ptr_data, type_data, data)  # return `data` to avoid the temporary copy is freed
wxchan's avatar
wxchan committed
649

wxchan's avatar
wxchan committed
650

651
def _is_allowed_numpy_dtype(dtype: type) -> bool:
652
    float128 = getattr(np, 'float128', type(None))
653
654
655
656
    return (
        issubclass(dtype, (np.integer, np.floating, np.bool_))
        and not issubclass(dtype, (np.timedelta64, float128))
    )
657
658


659
def _check_for_bad_pandas_dtypes(pandas_dtypes_series: pd_Series) -> None:
660
661
    bad_pandas_dtypes = [
        f'{column_name}: {pandas_dtype}'
662
        for column_name, pandas_dtype in pandas_dtypes_series.items()
663
        if not _is_allowed_numpy_dtype(pandas_dtype.type)
664
665
666
667
    ]
    if bad_pandas_dtypes:
        raise ValueError('pandas dtypes must be int, float or bool.\n'
                         f'Fields with bad pandas dtypes: {", ".join(bad_pandas_dtypes)}')
668
669


670
671
672
673
674
675
def _data_from_pandas(
    data,
    feature_name: Optional[_LGBM_FeatureNameConfiguration],
    categorical_feature: Optional[_LGBM_CategoricalFeatureConfiguration],
    pandas_categorical: Optional[List[List]]
):
676
    if isinstance(data, pd_DataFrame):
677
678
        if len(data.shape) != 2 or data.shape[0] < 1:
            raise ValueError('Input data must be 2 dimensional and non empty.')
679
        if feature_name == 'auto' or feature_name is None:
680
            data = data.rename(columns=str, copy=False)
681
        cat_cols = [col for col, dtype in zip(data.columns, data.dtypes) if isinstance(dtype, pd_CategoricalDtype)]
682
        cat_cols_not_ordered = [col for col in cat_cols if not data[col].cat.ordered]
683
684
685
686
687
        if pandas_categorical is None:  # train dataset
            pandas_categorical = [list(data[col].cat.categories) for col in cat_cols]
        else:
            if len(cat_cols) != len(pandas_categorical):
                raise ValueError('train and valid dataset categorical_feature do not match.')
688
            for col, category in zip(cat_cols, pandas_categorical):
689
690
                if list(data[col].cat.categories) != list(category):
                    data[col] = data[col].cat.set_categories(category)
691
        if len(cat_cols):  # cat_cols is list
692
            data = data.copy(deep=False)  # not alter origin DataFrame
693
            data[cat_cols] = data[cat_cols].apply(lambda x: x.cat.codes).replace({-1: np.nan})
694
695
696
        if categorical_feature is not None:
            if feature_name is None:
                feature_name = list(data.columns)
697
            if categorical_feature == 'auto':  # use cat cols from DataFrame
698
                categorical_feature = cat_cols_not_ordered
699
            else:  # use cat cols specified by user
700
                categorical_feature = list(categorical_feature)  # type: ignore[assignment]
701
702
        if feature_name == 'auto':
            feature_name = list(data.columns)
703
        _check_for_bad_pandas_dtypes(data.dtypes)
704
705
        df_dtypes = [dtype.type for dtype in data.dtypes]
        df_dtypes.append(np.float32)  # so that the target dtype considers floats
706
        target_dtype = np.result_type(*df_dtypes)
707
708
709
710
711
712
713
714
715
716
        try:
            # most common case (no nullable dtypes)
            data = data.to_numpy(dtype=target_dtype, copy=False)
        except TypeError:
            # 1.0 <= pd version < 1.1 and nullable dtypes, least common case
            # raises error because array is casted to type(pd.NA) and there's no na_value argument
            data = data.astype(target_dtype, copy=False).values
        except ValueError:
            # data has nullable dtypes, but we can specify na_value argument and copy will be made
            data = data.to_numpy(dtype=target_dtype, na_value=np.nan)
717
718
719
720
721
722
    else:
        if feature_name == 'auto':
            feature_name = None
        if categorical_feature == 'auto':
            categorical_feature = None
    return data, feature_name, categorical_feature, pandas_categorical
723
724


725
726
727
728
def _dump_pandas_categorical(
    pandas_categorical: Optional[List[List]],
    file_name: Optional[Union[str, Path]] = None
) -> str:
729
    categorical_json = json.dumps(pandas_categorical, default=_json_default_with_numpy)
730
    pandas_str = f'\npandas_categorical:{categorical_json}\n'
731
732
733
734
735
736
    if file_name is not None:
        with open(file_name, 'a') as f:
            f.write(pandas_str)
    return pandas_str


737
738
739
def _load_pandas_categorical(
    file_name: Optional[Union[str, Path]] = None,
    model_str: Optional[str] = None
740
) -> Optional[List[List]]:
741
742
    pandas_key = 'pandas_categorical:'
    offset = -len(pandas_key)
743
    if file_name is not None:
744
        max_offset = -getsize(file_name)
745
746
747
748
        with open(file_name, 'rb') as f:
            while True:
                if offset < max_offset:
                    offset = max_offset
749
                f.seek(offset, SEEK_END)
750
751
752
753
                lines = f.readlines()
                if len(lines) >= 2:
                    break
                offset *= 2
754
        last_line = lines[-1].decode('utf-8').strip()
755
        if not last_line.startswith(pandas_key):
756
            last_line = lines[-2].decode('utf-8').strip()
757
    elif model_str is not None:
758
759
760
761
762
763
        idx = model_str.rfind('\n', 0, offset)
        last_line = model_str[idx:].strip()
    if last_line.startswith(pandas_key):
        return json.loads(last_line[len(pandas_key):])
    else:
        return None
764
765


766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
class Sequence(abc.ABC):
    """
    Generic data access interface.

    Object should support the following operations:

    .. code-block::

        # Get total row number.
        >>> len(seq)
        # Random access by row index. Used for data sampling.
        >>> seq[10]
        # Range data access. Used to read data in batch when constructing Dataset.
        >>> seq[0:100]
        # Optionally specify batch_size to control range data read size.
        >>> seq.batch_size

    - With random access, **data sampling does not need to go through all data**.
    - With range data access, there's **no need to read all data into memory thus reduce memory usage**.

786
787
    .. versionadded:: 3.3.0

788
789
790
791
792
793
794
795
796
    Attributes
    ----------
    batch_size : int
        Default size of a batch.
    """

    batch_size = 4096  # Defaults to read 4K rows in each batch.

    @abc.abstractmethod
797
    def __getitem__(self, idx: Union[int, slice, List[int]]) -> np.ndarray:
798
799
800
801
802
803
804
        """Return data for given row index.

        A basic implementation should look like this:

        .. code-block:: python

            if isinstance(idx, numbers.Integral):
805
                return self._get_one_line(idx)
806
            elif isinstance(idx, slice):
807
808
                return np.stack([self._get_one_line(i) for i in range(idx.start, idx.stop)])
            elif isinstance(idx, list):
809
                # Only required if using ``Dataset.subset()``.
810
                return np.array([self._get_one_line(i) for i in idx])
811
            else:
812
                raise TypeError(f"Sequence index must be integer, slice or list, got {type(idx).__name__}")
813
814
815

        Parameters
        ----------
816
        idx : int, slice[int], list[int]
817
818
819
820
            Item index.

        Returns
        -------
821
        result : numpy 1-D array or numpy 2-D array
822
            1-D array if idx is int, 2-D array if idx is slice or list.
823
824
825
826
827
828
829
830
831
        """
        raise NotImplementedError("Sub-classes of lightgbm.Sequence must implement __getitem__()")

    @abc.abstractmethod
    def __len__(self) -> int:
        """Return row count of this sequence."""
        raise NotImplementedError("Sub-classes of lightgbm.Sequence must implement __len__()")


832
class _InnerPredictor:
833
834
835
836
837
    """_InnerPredictor of LightGBM.

    Not exposed to user.
    Used only for prediction, usually used for continued training.

Nikita Titov's avatar
Nikita Titov committed
838
839
840
    .. note::

        Can be converted from Booster, but cannot be converted to Booster.
Guolin Ke's avatar
Guolin Ke committed
841
    """
842

843
844
    def __init__(
        self,
845
846
847
848
        booster_handle: _BoosterHandle,
        pandas_categorical: Optional[List[List]],
        pred_parameter: Dict[str, Any],
        manage_handle: bool
849
    ):
850
        """Initialize the _InnerPredictor.
wxchan's avatar
wxchan committed
851
852
853

        Parameters
        ----------
854
        booster_handle : object
855
            Handle of Booster.
856
857
858
859
        pandas_categorical : list of list, or None
            If provided, list of categories for ``pandas`` categorical columns.
            Where the ``i``th element of the list contains the categories for the ``i``th categorical feature.
        pred_parameter : dict
860
            Other parameters for the prediction.
861
862
        manage_handle : bool
            If ``True``, free the corresponding Booster on the C++ side when this Python object is deleted.
wxchan's avatar
wxchan committed
863
        """
864
865
866
867
868
869
870
871
        self._handle = booster_handle
        self.__is_manage_handle = manage_handle
        self.pandas_categorical = pandas_categorical
        self.pred_parameter = _param_dict_to_str(pred_parameter)

        out_num_class = ctypes.c_int(0)
        _safe_call(
            _LIB.LGBM_BoosterGetNumClasses(
872
                self._handle,
873
874
875
876
                ctypes.byref(out_num_class)
            )
        )
        self.num_class = out_num_class.value
wxchan's avatar
wxchan committed
877

878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
    @classmethod
    def from_booster(
        cls,
        booster: "Booster",
        pred_parameter: Dict[str, Any]
    ) -> "_InnerPredictor":
        """Initialize an ``_InnerPredictor`` from a ``Booster``.

        Parameters
        ----------
        booster : Booster
            Booster.
        pred_parameter : dict
            Other parameters for the prediction.
        """
        out_cur_iter = ctypes.c_int(0)
        _safe_call(
            _LIB.LGBM_BoosterGetCurrentIteration(
                booster._handle,
                ctypes.byref(out_cur_iter)
            )
        )
        return cls(
            booster_handle=booster._handle,
            pandas_categorical=booster.pandas_categorical,
            pred_parameter=pred_parameter,
            manage_handle=False
        )

    @classmethod
    def from_model_file(
        cls,
        model_file: Union[str, Path],
        pred_parameter: Dict[str, Any]
    ) -> "_InnerPredictor":
        """Initialize an ``_InnerPredictor`` from a text file containing a LightGBM model.

        Parameters
        ----------
        model_file : str or pathlib.Path
            Path to the model file.
        pred_parameter : dict
            Other parameters for the prediction.
        """
        booster_handle = ctypes.c_void_p()
        out_num_iterations = ctypes.c_int(0)
        _safe_call(
            _LIB.LGBM_BoosterCreateFromModelfile(
                _c_str(str(model_file)),
                ctypes.byref(out_num_iterations),
                ctypes.byref(booster_handle)
            )
        )
        return cls(
            booster_handle=booster_handle,
            pandas_categorical=_load_pandas_categorical(file_name=model_file),
            pred_parameter=pred_parameter,
            manage_handle=True
        )
cbecker's avatar
cbecker committed
937

938
    def __del__(self) -> None:
939
940
        try:
            if self.__is_manage_handle:
941
                _safe_call(_LIB.LGBM_BoosterFree(self._handle))
942
943
        except AttributeError:
            pass
wxchan's avatar
wxchan committed
944

945
    def __getstate__(self) -> Dict[str, Any]:
946
947
        this = self.__dict__.copy()
        this.pop('handle', None)
948
        this.pop('_handle', None)
949
950
        return this

951
952
    def predict(
        self,
953
        data: _LGBM_PredictDataType,
954
955
956
957
958
959
960
        start_iteration: int = 0,
        num_iteration: int = -1,
        raw_score: bool = False,
        pred_leaf: bool = False,
        pred_contrib: bool = False,
        data_has_header: bool = False,
        validate_features: bool = False
961
    ) -> Union[np.ndarray, scipy.sparse.spmatrix, List[scipy.sparse.spmatrix]]:
962
        """Predict logic.
wxchan's avatar
wxchan committed
963
964
965

        Parameters
        ----------
966
        data : str, pathlib.Path, numpy array, pandas DataFrame, H2O DataTable's Frame or scipy.sparse
967
            Data source for prediction.
968
            If str or pathlib.Path, it represents the path to a text file (CSV, TSV, or LibSVM).
969
970
        start_iteration : int, optional (default=0)
            Start index of the iteration to predict.
971
972
973
974
975
976
977
978
979
980
981
        num_iteration : int, optional (default=-1)
            Iteration used for prediction.
        raw_score : bool, optional (default=False)
            Whether to predict raw scores.
        pred_leaf : bool, optional (default=False)
            Whether to predict leaf index.
        pred_contrib : bool, optional (default=False)
            Whether to predict feature contributions.
        data_has_header : bool, optional (default=False)
            Whether data has header.
            Used only for txt data.
982
983
984
        validate_features : bool, optional (default=False)
            If True, ensure that the features used to predict match the ones used to train.
            Used only if data is pandas DataFrame.
wxchan's avatar
wxchan committed
985

986
987
            .. versionadded:: 4.0.0

wxchan's avatar
wxchan committed
988
989
        Returns
        -------
990
        result : numpy array, scipy.sparse or list of scipy.sparse
991
            Prediction result.
992
            Can be sparse or a list of sparse objects (each element represents predictions for one class) for feature contributions (when ``pred_contrib=True``).
wxchan's avatar
wxchan committed
993
        """
wxchan's avatar
wxchan committed
994
        if isinstance(data, Dataset):
995
            raise TypeError("Cannot use Dataset instance for prediction, please use raw data instead")
996
997
998
999
1000
1001
        elif isinstance(data, pd_DataFrame) and validate_features:
            data_names = [str(x) for x in data.columns]
            ptr_names = (ctypes.c_char_p * len(data_names))()
            ptr_names[:] = [x.encode('utf-8') for x in data_names]
            _safe_call(
                _LIB.LGBM_BoosterValidateFeatureNames(
1002
                    self._handle,
1003
1004
1005
1006
                    ptr_names,
                    ctypes.c_int(len(data_names)),
                )
            )
1007
        data = _data_from_pandas(data, None, None, self.pandas_categorical)[0]
1008
        predict_type = _C_API_PREDICT_NORMAL
wxchan's avatar
wxchan committed
1009
        if raw_score:
1010
            predict_type = _C_API_PREDICT_RAW_SCORE
wxchan's avatar
wxchan committed
1011
        if pred_leaf:
1012
            predict_type = _C_API_PREDICT_LEAF_INDEX
1013
        if pred_contrib:
1014
            predict_type = _C_API_PREDICT_CONTRIB
cbecker's avatar
cbecker committed
1015

1016
        if isinstance(data, (str, Path)):
1017
            with _TempFile() as f:
wxchan's avatar
wxchan committed
1018
                _safe_call(_LIB.LGBM_BoosterPredictForFile(
1019
                    self._handle,
1020
                    _c_str(str(data)),
1021
                    ctypes.c_int(data_has_header),
Guolin Ke's avatar
Guolin Ke committed
1022
                    ctypes.c_int(predict_type),
1023
                    ctypes.c_int(start_iteration),
Guolin Ke's avatar
Guolin Ke committed
1024
                    ctypes.c_int(num_iteration),
1025
1026
                    _c_str(self.pred_parameter),
                    _c_str(f.name)))
1027
1028
                preds = np.loadtxt(f.name, dtype=np.float64)
                nrow = preds.shape[0]
wxchan's avatar
wxchan committed
1029
        elif isinstance(data, scipy.sparse.csr_matrix):
1030
1031
1032
1033
1034
1035
            preds, nrow = self.__pred_for_csr(
                csr=data,
                start_iteration=start_iteration,
                num_iteration=num_iteration,
                predict_type=predict_type
            )
Guolin Ke's avatar
Guolin Ke committed
1036
        elif isinstance(data, scipy.sparse.csc_matrix):
1037
1038
1039
1040
1041
1042
            preds, nrow = self.__pred_for_csc(
                csc=data,
                start_iteration=start_iteration,
                num_iteration=num_iteration,
                predict_type=predict_type
            )
wxchan's avatar
wxchan committed
1043
        elif isinstance(data, np.ndarray):
1044
1045
1046
1047
1048
1049
            preds, nrow = self.__pred_for_np2d(
                mat=data,
                start_iteration=start_iteration,
                num_iteration=num_iteration,
                predict_type=predict_type
            )
1050
1051
1052
        elif isinstance(data, list):
            try:
                data = np.array(data)
1053
1054
            except BaseException as err:
                raise ValueError('Cannot convert data list to numpy array.') from err
1055
1056
1057
1058
1059
1060
            preds, nrow = self.__pred_for_np2d(
                mat=data,
                start_iteration=start_iteration,
                num_iteration=num_iteration,
                predict_type=predict_type
            )
1061
        elif isinstance(data, dt_DataTable):
1062
1063
1064
1065
1066
1067
            preds, nrow = self.__pred_for_np2d(
                mat=data.to_numpy(),
                start_iteration=start_iteration,
                num_iteration=num_iteration,
                predict_type=predict_type
            )
wxchan's avatar
wxchan committed
1068
1069
        else:
            try:
1070
                _log_warning('Converting data to scipy sparse matrix.')
wxchan's avatar
wxchan committed
1071
                csr = scipy.sparse.csr_matrix(data)
1072
1073
            except BaseException as err:
                raise TypeError(f'Cannot predict data for type {type(data).__name__}') from err
1074
1075
1076
1077
1078
1079
            preds, nrow = self.__pred_for_csr(
                csr=csr,
                start_iteration=start_iteration,
                num_iteration=num_iteration,
                predict_type=predict_type
            )
wxchan's avatar
wxchan committed
1080
1081
        if pred_leaf:
            preds = preds.astype(np.int32)
1082
        is_sparse = isinstance(preds, scipy.sparse.spmatrix) or isinstance(preds, list)
1083
        if not is_sparse and preds.size != nrow:
wxchan's avatar
wxchan committed
1084
            if preds.size % nrow == 0:
1085
                preds = preds.reshape(nrow, -1)
wxchan's avatar
wxchan committed
1086
            else:
1087
                raise ValueError(f'Length of predict result ({preds.size}) cannot be divide nrow ({nrow})')
wxchan's avatar
wxchan committed
1088
1089
        return preds

1090
1091
1092
1093
1094
1095
1096
    def __get_num_preds(
        self,
        start_iteration: int,
        num_iteration: int,
        nrow: int,
        predict_type: int
    ) -> int:
1097
        """Get size of prediction result."""
1098
        if nrow > _MAX_INT32:
1099
            raise LightGBMError('LightGBM cannot perform prediction for data '
1100
                                f'with number of rows greater than MAX_INT32 ({_MAX_INT32}).\n'
1101
                                'You can split your data into chunks '
1102
                                'and then concatenate predictions for them')
Guolin Ke's avatar
Guolin Ke committed
1103
1104
        n_preds = ctypes.c_int64(0)
        _safe_call(_LIB.LGBM_BoosterCalcNumPredict(
1105
            self._handle,
Guolin Ke's avatar
Guolin Ke committed
1106
1107
            ctypes.c_int(nrow),
            ctypes.c_int(predict_type),
1108
            ctypes.c_int(start_iteration),
Guolin Ke's avatar
Guolin Ke committed
1109
            ctypes.c_int(num_iteration),
Guolin Ke's avatar
Guolin Ke committed
1110
1111
            ctypes.byref(n_preds)))
        return n_preds.value
wxchan's avatar
wxchan committed
1112

1113
1114
1115
1116
1117
1118
1119
1120
1121
1122
1123
1124
1125
    def __inner_predict_np2d(
        self,
        mat: np.ndarray,
        start_iteration: int,
        num_iteration: int,
        predict_type: int,
        preds: Optional[np.ndarray]
    ) -> Tuple[np.ndarray, int]:
        if mat.dtype == np.float32 or mat.dtype == np.float64:
            data = np.array(mat.reshape(mat.size), dtype=mat.dtype, copy=False)
        else:  # change non-float data to float data, need to copy
            data = np.array(mat.reshape(mat.size), dtype=np.float32)
        ptr_data, type_ptr_data, _ = _c_float_array(data)
1126
1127
1128
1129
1130
1131
        n_preds = self.__get_num_preds(
            start_iteration=start_iteration,
            num_iteration=num_iteration,
            nrow=mat.shape[0],
            predict_type=predict_type
        )
1132
1133
1134
1135
1136
1137
        if preds is None:
            preds = np.empty(n_preds, dtype=np.float64)
        elif len(preds.shape) != 1 or len(preds) != n_preds:
            raise ValueError("Wrong length of pre-allocated predict array")
        out_num_preds = ctypes.c_int64(0)
        _safe_call(_LIB.LGBM_BoosterPredictForMat(
1138
            self._handle,
1139
1140
1141
1142
1143
1144
1145
1146
1147
1148
1149
1150
1151
1152
1153
1154
1155
1156
1157
1158
1159
1160
            ptr_data,
            ctypes.c_int(type_ptr_data),
            ctypes.c_int32(mat.shape[0]),
            ctypes.c_int32(mat.shape[1]),
            ctypes.c_int(_C_API_IS_ROW_MAJOR),
            ctypes.c_int(predict_type),
            ctypes.c_int(start_iteration),
            ctypes.c_int(num_iteration),
            _c_str(self.pred_parameter),
            ctypes.byref(out_num_preds),
            preds.ctypes.data_as(ctypes.POINTER(ctypes.c_double))))
        if n_preds != out_num_preds.value:
            raise ValueError("Wrong length for predict results")
        return preds, mat.shape[0]

    def __pred_for_np2d(
        self,
        mat: np.ndarray,
        start_iteration: int,
        num_iteration: int,
        predict_type: int
    ) -> Tuple[np.ndarray, int]:
1161
        """Predict for a 2-D numpy matrix."""
wxchan's avatar
wxchan committed
1162
        if len(mat.shape) != 2:
1163
            raise ValueError('Input numpy.ndarray or list must be 2 dimensional')
wxchan's avatar
wxchan committed
1164

1165
        nrow = mat.shape[0]
1166
1167
        if nrow > _MAX_INT32:
            sections = np.arange(start=_MAX_INT32, stop=nrow, step=_MAX_INT32)
1168
            # __get_num_preds() cannot work with nrow > MAX_INT32, so calculate overall number of predictions piecemeal
1169
            n_preds = [self.__get_num_preds(start_iteration, num_iteration, i, predict_type) for i in np.diff([0] + list(sections) + [nrow])]
1170
            n_preds_sections = np.array([0] + n_preds, dtype=np.intp).cumsum()
1171
            preds = np.empty(sum(n_preds), dtype=np.float64)
1172
1173
            for chunk, (start_idx_pred, end_idx_pred) in zip(np.array_split(mat, sections),
                                                             zip(n_preds_sections, n_preds_sections[1:])):
1174
                # avoid memory consumption by arrays concatenation operations
1175
1176
1177
1178
1179
1180
1181
                self.__inner_predict_np2d(
                    mat=chunk,
                    start_iteration=start_iteration,
                    num_iteration=num_iteration,
                    predict_type=predict_type,
                    preds=preds[start_idx_pred:end_idx_pred]
                )
1182
            return preds, nrow
wxchan's avatar
wxchan committed
1183
        else:
1184
1185
1186
1187
1188
1189
1190
            return self.__inner_predict_np2d(
                mat=mat,
                start_iteration=start_iteration,
                num_iteration=num_iteration,
                predict_type=predict_type,
                preds=None
            )
wxchan's avatar
wxchan committed
1191

1192
1193
1194
    def __create_sparse_native(
        self,
        cs: Union[scipy.sparse.csc_matrix, scipy.sparse.csr_matrix],
1195
1196
1197
1198
1199
1200
        out_shape: np.ndarray,
        out_ptr_indptr: "ctypes._Pointer",
        out_ptr_indices: "ctypes._Pointer",
        out_ptr_data: "ctypes._Pointer",
        indptr_type: int,
        data_type: int,
1201
        is_csr: bool
1202
    ) -> Union[List[scipy.sparse.csc_matrix], List[scipy.sparse.csr_matrix]]:
1203
1204
1205
        # create numpy array from output arrays
        data_indices_len = out_shape[0]
        indptr_len = out_shape[1]
1206
        if indptr_type == _C_API_DTYPE_INT32:
1207
            out_indptr = _cint32_array_to_numpy(out_ptr_indptr, indptr_len)
1208
        elif indptr_type == _C_API_DTYPE_INT64:
1209
            out_indptr = _cint64_array_to_numpy(out_ptr_indptr, indptr_len)
1210
1211
        else:
            raise TypeError("Expected int32 or int64 type for indptr")
1212
        if data_type == _C_API_DTYPE_FLOAT32:
1213
            out_data = _cfloat32_array_to_numpy(out_ptr_data, data_indices_len)
1214
        elif data_type == _C_API_DTYPE_FLOAT64:
1215
            out_data = _cfloat64_array_to_numpy(out_ptr_data, data_indices_len)
1216
1217
        else:
            raise TypeError("Expected float32 or float64 type for data")
1218
        out_indices = _cint32_array_to_numpy(out_ptr_indices, data_indices_len)
1219
1220
1221
1222
1223
1224
1225
1226
1227
1228
1229
1230
1231
1232
1233
1234
1235
1236
1237
1238
1239
1240
1241
1242
1243
1244
1245
1246
        # break up indptr based on number of rows (note more than one matrix in multiclass case)
        per_class_indptr_shape = cs.indptr.shape[0]
        # for CSC there is extra column added
        if not is_csr:
            per_class_indptr_shape += 1
        out_indptr_arrays = np.split(out_indptr, out_indptr.shape[0] / per_class_indptr_shape)
        # reformat output into a csr or csc matrix or list of csr or csc matrices
        cs_output_matrices = []
        offset = 0
        for cs_indptr in out_indptr_arrays:
            matrix_indptr_len = cs_indptr[cs_indptr.shape[0] - 1]
            cs_indices = out_indices[offset + cs_indptr[0]:offset + matrix_indptr_len]
            cs_data = out_data[offset + cs_indptr[0]:offset + matrix_indptr_len]
            offset += matrix_indptr_len
            # same shape as input csr or csc matrix except extra column for expected value
            cs_shape = [cs.shape[0], cs.shape[1] + 1]
            # note: make sure we copy data as it will be deallocated next
            if is_csr:
                cs_output_matrices.append(scipy.sparse.csr_matrix((cs_data, cs_indices, cs_indptr), cs_shape))
            else:
                cs_output_matrices.append(scipy.sparse.csc_matrix((cs_data, cs_indices, cs_indptr), cs_shape))
        # free the temporary native indptr, indices, and data
        _safe_call(_LIB.LGBM_BoosterFreePredictSparse(out_ptr_indptr, out_ptr_indices, out_ptr_data,
                                                      ctypes.c_int(indptr_type), ctypes.c_int(data_type)))
        if len(cs_output_matrices) == 1:
            return cs_output_matrices[0]
        return cs_output_matrices

1247
1248
1249
1250
1251
1252
1253
1254
1255
    def __inner_predict_csr(
        self,
        csr: scipy.sparse.csr_matrix,
        start_iteration: int,
        num_iteration: int,
        predict_type: int,
        preds: Optional[np.ndarray]
    ) -> Tuple[np.ndarray, int]:
        nrow = len(csr.indptr) - 1
1256
1257
1258
1259
1260
1261
        n_preds = self.__get_num_preds(
            start_iteration=start_iteration,
            num_iteration=num_iteration,
            nrow=nrow,
            predict_type=predict_type
        )
1262
1263
1264
1265
1266
        if preds is None:
            preds = np.empty(n_preds, dtype=np.float64)
        elif len(preds.shape) != 1 or len(preds) != n_preds:
            raise ValueError("Wrong length of pre-allocated predict array")
        out_num_preds = ctypes.c_int64(0)
wxchan's avatar
wxchan committed
1267

1268
1269
        ptr_indptr, type_ptr_indptr, _ = _c_int_array(csr.indptr)
        ptr_data, type_ptr_data, _ = _c_float_array(csr.data)
1270

1271
1272
1273
1274
        assert csr.shape[1] <= _MAX_INT32
        csr_indices = csr.indices.astype(np.int32, copy=False)

        _safe_call(_LIB.LGBM_BoosterPredictForCSR(
1275
            self._handle,
1276
1277
1278
1279
1280
1281
1282
1283
1284
1285
1286
1287
1288
1289
1290
1291
1292
1293
1294
1295
1296
1297
1298
1299
            ptr_indptr,
            ctypes.c_int(type_ptr_indptr),
            csr_indices.ctypes.data_as(ctypes.POINTER(ctypes.c_int32)),
            ptr_data,
            ctypes.c_int(type_ptr_data),
            ctypes.c_int64(len(csr.indptr)),
            ctypes.c_int64(len(csr.data)),
            ctypes.c_int64(csr.shape[1]),
            ctypes.c_int(predict_type),
            ctypes.c_int(start_iteration),
            ctypes.c_int(num_iteration),
            _c_str(self.pred_parameter),
            ctypes.byref(out_num_preds),
            preds.ctypes.data_as(ctypes.POINTER(ctypes.c_double))))
        if n_preds != out_num_preds.value:
            raise ValueError("Wrong length for predict results")
        return preds, nrow

    def __inner_predict_csr_sparse(
        self,
        csr: scipy.sparse.csr_matrix,
        start_iteration: int,
        num_iteration: int,
        predict_type: int
1300
    ) -> Tuple[Union[List[scipy.sparse.csc_matrix], List[scipy.sparse.csr_matrix]], int]:
1301
1302
1303
1304
        ptr_indptr, type_ptr_indptr, __ = _c_int_array(csr.indptr)
        ptr_data, type_ptr_data, _ = _c_float_array(csr.data)
        csr_indices = csr.indices.astype(np.int32, copy=False)
        matrix_type = _C_API_MATRIX_TYPE_CSR
1305
        out_ptr_indptr: _ctypes_int_ptr
1306
1307
1308
1309
1310
        if type_ptr_indptr == _C_API_DTYPE_INT32:
            out_ptr_indptr = ctypes.POINTER(ctypes.c_int32)()
        else:
            out_ptr_indptr = ctypes.POINTER(ctypes.c_int64)()
        out_ptr_indices = ctypes.POINTER(ctypes.c_int32)()
1311
        out_ptr_data: _ctypes_float_ptr
1312
1313
1314
1315
1316
1317
        if type_ptr_data == _C_API_DTYPE_FLOAT32:
            out_ptr_data = ctypes.POINTER(ctypes.c_float)()
        else:
            out_ptr_data = ctypes.POINTER(ctypes.c_double)()
        out_shape = np.empty(2, dtype=np.int64)
        _safe_call(_LIB.LGBM_BoosterPredictSparseOutput(
1318
            self._handle,
1319
1320
1321
1322
1323
1324
1325
1326
1327
1328
1329
1330
1331
1332
1333
1334
1335
1336
1337
1338
1339
1340
1341
1342
1343
1344
1345
1346
1347
1348
            ptr_indptr,
            ctypes.c_int(type_ptr_indptr),
            csr_indices.ctypes.data_as(ctypes.POINTER(ctypes.c_int32)),
            ptr_data,
            ctypes.c_int(type_ptr_data),
            ctypes.c_int64(len(csr.indptr)),
            ctypes.c_int64(len(csr.data)),
            ctypes.c_int64(csr.shape[1]),
            ctypes.c_int(predict_type),
            ctypes.c_int(start_iteration),
            ctypes.c_int(num_iteration),
            _c_str(self.pred_parameter),
            ctypes.c_int(matrix_type),
            out_shape.ctypes.data_as(ctypes.POINTER(ctypes.c_int64)),
            ctypes.byref(out_ptr_indptr),
            ctypes.byref(out_ptr_indices),
            ctypes.byref(out_ptr_data)))
        matrices = self.__create_sparse_native(
            cs=csr,
            out_shape=out_shape,
            out_ptr_indptr=out_ptr_indptr,
            out_ptr_indices=out_ptr_indices,
            out_ptr_data=out_ptr_data,
            indptr_type=type_ptr_indptr,
            data_type=type_ptr_data,
            is_csr=True
        )
        nrow = len(csr.indptr) - 1
        return matrices, nrow

1349
1350
1351
1352
1353
1354
1355
    def __pred_for_csr(
        self,
        csr: scipy.sparse.csr_matrix,
        start_iteration: int,
        num_iteration: int,
        predict_type: int
    ) -> Tuple[np.ndarray, int]:
1356
        """Predict for a CSR data."""
1357
        if predict_type == _C_API_PREDICT_CONTRIB:
1358
1359
1360
1361
1362
1363
            return self.__inner_predict_csr_sparse(
                csr=csr,
                start_iteration=start_iteration,
                num_iteration=num_iteration,
                predict_type=predict_type
            )
1364
        nrow = len(csr.indptr) - 1
1365
1366
        if nrow > _MAX_INT32:
            sections = [0] + list(np.arange(start=_MAX_INT32, stop=nrow, step=_MAX_INT32)) + [nrow]
1367
            # __get_num_preds() cannot work with nrow > MAX_INT32, so calculate overall number of predictions piecemeal
1368
            n_preds = [self.__get_num_preds(start_iteration, num_iteration, i, predict_type) for i in np.diff(sections)]
1369
            n_preds_sections = np.array([0] + n_preds, dtype=np.intp).cumsum()
1370
            preds = np.empty(sum(n_preds), dtype=np.float64)
1371
1372
            for (start_idx, end_idx), (start_idx_pred, end_idx_pred) in zip(zip(sections, sections[1:]),
                                                                            zip(n_preds_sections, n_preds_sections[1:])):
1373
                # avoid memory consumption by arrays concatenation operations
1374
1375
1376
1377
1378
1379
1380
                self.__inner_predict_csr(
                    csr=csr[start_idx:end_idx],
                    start_iteration=start_iteration,
                    num_iteration=num_iteration,
                    predict_type=predict_type,
                    preds=preds[start_idx_pred:end_idx_pred]
                )
1381
1382
            return preds, nrow
        else:
1383
1384
1385
1386
1387
1388
1389
1390
1391
1392
            return self.__inner_predict_csr(
                csr=csr,
                start_iteration=start_iteration,
                num_iteration=num_iteration,
                predict_type=predict_type,
                preds=None
            )

    def __inner_predict_sparse_csc(
        self,
1393
1394
1395
1396
        csc: scipy.sparse.csc_matrix,
        start_iteration: int,
        num_iteration: int,
        predict_type: int
1397
1398
1399
1400
1401
    ):
        ptr_indptr, type_ptr_indptr, __ = _c_int_array(csc.indptr)
        ptr_data, type_ptr_data, _ = _c_float_array(csc.data)
        csc_indices = csc.indices.astype(np.int32, copy=False)
        matrix_type = _C_API_MATRIX_TYPE_CSC
1402
        out_ptr_indptr: _ctypes_int_ptr
1403
1404
1405
1406
1407
        if type_ptr_indptr == _C_API_DTYPE_INT32:
            out_ptr_indptr = ctypes.POINTER(ctypes.c_int32)()
        else:
            out_ptr_indptr = ctypes.POINTER(ctypes.c_int64)()
        out_ptr_indices = ctypes.POINTER(ctypes.c_int32)()
1408
        out_ptr_data: _ctypes_float_ptr
1409
1410
1411
1412
1413
1414
        if type_ptr_data == _C_API_DTYPE_FLOAT32:
            out_ptr_data = ctypes.POINTER(ctypes.c_float)()
        else:
            out_ptr_data = ctypes.POINTER(ctypes.c_double)()
        out_shape = np.empty(2, dtype=np.int64)
        _safe_call(_LIB.LGBM_BoosterPredictSparseOutput(
1415
            self._handle,
1416
1417
1418
1419
1420
1421
1422
1423
1424
1425
1426
1427
1428
1429
1430
1431
1432
1433
1434
1435
1436
1437
1438
1439
1440
1441
1442
1443
1444
            ptr_indptr,
            ctypes.c_int(type_ptr_indptr),
            csc_indices.ctypes.data_as(ctypes.POINTER(ctypes.c_int32)),
            ptr_data,
            ctypes.c_int(type_ptr_data),
            ctypes.c_int64(len(csc.indptr)),
            ctypes.c_int64(len(csc.data)),
            ctypes.c_int64(csc.shape[0]),
            ctypes.c_int(predict_type),
            ctypes.c_int(start_iteration),
            ctypes.c_int(num_iteration),
            _c_str(self.pred_parameter),
            ctypes.c_int(matrix_type),
            out_shape.ctypes.data_as(ctypes.POINTER(ctypes.c_int64)),
            ctypes.byref(out_ptr_indptr),
            ctypes.byref(out_ptr_indices),
            ctypes.byref(out_ptr_data)))
        matrices = self.__create_sparse_native(
            cs=csc,
            out_shape=out_shape,
            out_ptr_indptr=out_ptr_indptr,
            out_ptr_indices=out_ptr_indices,
            out_ptr_data=out_ptr_data,
            indptr_type=type_ptr_indptr,
            data_type=type_ptr_data,
            is_csr=False
        )
        nrow = csc.shape[0]
        return matrices, nrow
Guolin Ke's avatar
Guolin Ke committed
1445

1446
1447
1448
1449
1450
1451
1452
    def __pred_for_csc(
        self,
        csc: scipy.sparse.csc_matrix,
        start_iteration: int,
        num_iteration: int,
        predict_type: int
    ) -> Tuple[np.ndarray, int]:
1453
        """Predict for a CSC data."""
Guolin Ke's avatar
Guolin Ke committed
1454
        nrow = csc.shape[0]
1455
        if nrow > _MAX_INT32:
1456
1457
1458
1459
1460
1461
            return self.__pred_for_csr(
                csr=csc.tocsr(),
                start_iteration=start_iteration,
                num_iteration=num_iteration,
                predict_type=predict_type
            )
1462
        if predict_type == _C_API_PREDICT_CONTRIB:
1463
1464
1465
1466
1467
1468
            return self.__inner_predict_sparse_csc(
                csc=csc,
                start_iteration=start_iteration,
                num_iteration=num_iteration,
                predict_type=predict_type
            )
1469
1470
1471
1472
1473
1474
        n_preds = self.__get_num_preds(
            start_iteration=start_iteration,
            num_iteration=num_iteration,
            nrow=nrow,
            predict_type=predict_type
        )
1475
        preds = np.empty(n_preds, dtype=np.float64)
Guolin Ke's avatar
Guolin Ke committed
1476
1477
        out_num_preds = ctypes.c_int64(0)

1478
1479
        ptr_indptr, type_ptr_indptr, __ = _c_int_array(csc.indptr)
        ptr_data, type_ptr_data, _ = _c_float_array(csc.data)
Guolin Ke's avatar
Guolin Ke committed
1480

1481
        assert csc.shape[0] <= _MAX_INT32
1482
        csc_indices = csc.indices.astype(np.int32, copy=False)
1483

Guolin Ke's avatar
Guolin Ke committed
1484
        _safe_call(_LIB.LGBM_BoosterPredictForCSC(
1485
            self._handle,
Guolin Ke's avatar
Guolin Ke committed
1486
            ptr_indptr,
1487
            ctypes.c_int(type_ptr_indptr),
1488
            csc_indices.ctypes.data_as(ctypes.POINTER(ctypes.c_int32)),
Guolin Ke's avatar
Guolin Ke committed
1489
            ptr_data,
Guolin Ke's avatar
Guolin Ke committed
1490
1491
1492
1493
1494
            ctypes.c_int(type_ptr_data),
            ctypes.c_int64(len(csc.indptr)),
            ctypes.c_int64(len(csc.data)),
            ctypes.c_int64(csc.shape[0]),
            ctypes.c_int(predict_type),
1495
            ctypes.c_int(start_iteration),
Guolin Ke's avatar
Guolin Ke committed
1496
            ctypes.c_int(num_iteration),
1497
            _c_str(self.pred_parameter),
Guolin Ke's avatar
Guolin Ke committed
1498
            ctypes.byref(out_num_preds),
wxchan's avatar
wxchan committed
1499
            preds.ctypes.data_as(ctypes.POINTER(ctypes.c_double))))
wxchan's avatar
wxchan committed
1500
        if n_preds != out_num_preds.value:
1501
            raise ValueError("Wrong length for predict results")
wxchan's avatar
wxchan committed
1502
1503
        return preds, nrow

1504
    def current_iteration(self) -> int:
1505
1506
1507
1508
1509
1510
1511
1512
1513
        """Get the index of the current iteration.

        Returns
        -------
        cur_iter : int
            The index of the current iteration.
        """
        out_cur_iter = ctypes.c_int(0)
        _safe_call(_LIB.LGBM_BoosterGetCurrentIteration(
1514
            self._handle,
1515
1516
1517
            ctypes.byref(out_cur_iter)))
        return out_cur_iter.value

wxchan's avatar
wxchan committed
1518

1519
class Dataset:
wxchan's avatar
wxchan committed
1520
    """Dataset in LightGBM."""
1521

1522
1523
    def __init__(
        self,
1524
        data: _LGBM_TrainDataType,
1525
        label: Optional[_LGBM_LabelType] = None,
1526
        reference: Optional["Dataset"] = None,
1527
1528
1529
        weight: Optional[_LGBM_WeightType] = None,
        group: Optional[_LGBM_GroupType] = None,
        init_score: Optional[_LGBM_InitScoreType] = None,
1530
1531
        feature_name: _LGBM_FeatureNameConfiguration = 'auto',
        categorical_feature: _LGBM_CategoricalFeatureConfiguration = 'auto',
1532
        params: Optional[Dict[str, Any]] = None,
1533
1534
        free_raw_data: bool = True,
        position: Optional[_LGBM_PositionType] = None,
1535
    ):
1536
        """Initialize Dataset.
1537

wxchan's avatar
wxchan committed
1538
1539
        Parameters
        ----------
1540
        data : str, pathlib.Path, numpy array, pandas DataFrame, H2O DataTable's Frame, scipy.sparse, Sequence, list of Sequence or list of numpy array
wxchan's avatar
wxchan committed
1541
            Data source of Dataset.
1542
            If str or pathlib.Path, it represents the path to a text file (CSV, TSV, or LibSVM) or a LightGBM Dataset binary file.
1543
        label : list, numpy 1-D array, pandas Series / one-column DataFrame or None, optional (default=None)
1544
1545
1546
            Label of the data.
        reference : Dataset or None, optional (default=None)
            If this is Dataset for validation, training data should be used as reference.
1547
        weight : list, numpy 1-D array, pandas Series or None, optional (default=None)
1548
            Weight for each instance. Weights should be non-negative.
1549
        group : list, numpy 1-D array, pandas Series or None, optional (default=None)
1550
1551
1552
            Group/query data.
            Only used in the learning-to-rank task.
            sum(group) = n_samples.
1553
1554
            For example, if you have a 100-document dataset with ``group = [10, 20, 40, 10, 10, 10]``, that means that you have 6 groups,
            where the first 10 records are in the first group, records 11-30 are in the second group, records 31-70 are in the third group, etc.
1555
        init_score : list, list of lists (for multi-class task), numpy array, pandas Series, pandas DataFrame (for multi-class task), or None, optional (default=None)
1556
            Init score for Dataset.
1557
        feature_name : list of str, or 'auto', optional (default="auto")
1558
1559
            Feature names.
            If 'auto' and data is pandas DataFrame, data columns names are used.
1560
        categorical_feature : list of str or int, or 'auto', optional (default="auto")
1561
1562
            Categorical features.
            If list of int, interpreted as indices.
1563
            If list of str, interpreted as feature names (need to specify ``feature_name`` as well).
1564
            If 'auto' and data is pandas DataFrame, pandas unordered categorical columns are used.
1565
            All values in categorical features will be cast to int32 and thus should be less than int32 max value (2147483647).
1566
            Large values could be memory consuming. Consider using consecutive integers starting from zero.
1567
            All negative values in categorical features will be treated as missing values.
1568
            The output cannot be monotonically constrained with respect to a categorical feature.
1569
            Floating point numbers in categorical features will be rounded towards 0.
Nikita Titov's avatar
Nikita Titov committed
1570
        params : dict or None, optional (default=None)
1571
            Other parameters for Dataset.
Nikita Titov's avatar
Nikita Titov committed
1572
        free_raw_data : bool, optional (default=True)
1573
            If True, raw data is freed after constructing inner Dataset.
1574
1575
        position : numpy 1-D array, pandas Series or None, optional (default=None)
            Position of items used in unbiased learning-to-rank task.
wxchan's avatar
wxchan committed
1576
        """
1577
        self._handle: Optional[_DatasetHandle] = None
wxchan's avatar
wxchan committed
1578
1579
1580
1581
1582
        self.data = data
        self.label = label
        self.reference = reference
        self.weight = weight
        self.group = group
1583
        self.position = position
1584
        self.init_score = init_score
1585
1586
        self.feature_name: _LGBM_FeatureNameConfiguration = feature_name
        self.categorical_feature: _LGBM_CategoricalFeatureConfiguration = categorical_feature
1587
        self.params = deepcopy(params)
wxchan's avatar
wxchan committed
1588
        self.free_raw_data = free_raw_data
1589
        self.used_indices: Optional[List[int]] = None
1590
        self._need_slice = True
1591
        self._predictor: Optional[_InnerPredictor] = None
1592
        self.pandas_categorical = None
1593
        self._params_back_up = None
1594
        self.version = 0
1595
        self._start_row = 0  # Used when pushing rows one by one.
wxchan's avatar
wxchan committed
1596

1597
    def __del__(self) -> None:
1598
1599
1600
1601
        try:
            self._free_handle()
        except AttributeError:
            pass
1602

1603
1604
1605
1606
1607
1608
1609
1610
1611
1612
1613
1614
1615
1616
1617
1618
1619
    def _create_sample_indices(self, total_nrow: int) -> np.ndarray:
        """Get an array of randomly chosen indices from this ``Dataset``.

        Indices are sampled without replacement.

        Parameters
        ----------
        total_nrow : int
            Total number of rows to sample from.
            If this value is greater than the value of parameter ``bin_construct_sample_cnt``, only ``bin_construct_sample_cnt`` indices will be used.
            If Dataset has multiple input data, this should be the sum of rows of every file.

        Returns
        -------
        indices : numpy array
            Indices for sampled data.
        """
1620
        param_str = _param_dict_to_str(self.get_params())
1621
1622
        sample_cnt = _get_sample_count(total_nrow, param_str)
        indices = np.empty(sample_cnt, dtype=np.int32)
1623
        ptr_data, _, _ = _c_int_array(indices)
1624
1625
1626
1627
        actual_sample_cnt = ctypes.c_int32(0)

        _safe_call(_LIB.LGBM_SampleIndices(
            ctypes.c_int32(total_nrow),
1628
            _c_str(param_str),
1629
1630
1631
            ptr_data,
            ctypes.byref(actual_sample_cnt),
        ))
1632
1633
        assert sample_cnt == actual_sample_cnt.value
        return indices
1634

1635
1636
1637
1638
1639
    def _init_from_ref_dataset(
        self,
        total_nrow: int,
        ref_dataset: _DatasetHandle
    ) -> 'Dataset':
1640
1641
1642
1643
1644
1645
        """Create dataset from a reference dataset.

        Parameters
        ----------
        total_nrow : int
            Number of rows expected to add to dataset.
1646
1647
        ref_dataset : object
            Handle of reference dataset to extract metadata from.
1648
1649
1650
1651
1652
1653

        Returns
        -------
        self : Dataset
            Constructed Dataset object.
        """
1654
        self._handle = ctypes.c_void_p()
1655
1656
1657
        _safe_call(_LIB.LGBM_DatasetCreateByReference(
            ref_dataset,
            ctypes.c_int64(total_nrow),
1658
            ctypes.byref(self._handle),
1659
1660
1661
1662
1663
1664
1665
1666
1667
1668
1669
1670
1671
1672
        ))
        return self

    def _init_from_sample(
        self,
        sample_data: List[np.ndarray],
        sample_indices: List[np.ndarray],
        sample_cnt: int,
        total_nrow: int,
    ) -> "Dataset":
        """Create Dataset from sampled data structures.

        Parameters
        ----------
1673
        sample_data : list of numpy array
1674
            Sample data for each column.
1675
        sample_indices : list of numpy array
1676
1677
1678
1679
1680
1681
1682
1683
1684
1685
1686
1687
1688
1689
1690
1691
1692
1693
1694
1695
1696
1697
            Sample data row index for each column.
        sample_cnt : int
            Number of samples.
        total_nrow : int
            Total number of rows for all input files.

        Returns
        -------
        self : Dataset
            Constructed Dataset object.
        """
        ncol = len(sample_indices)
        assert len(sample_data) == ncol, "#sample data column != #column indices"

        for i in range(ncol):
            if sample_data[i].dtype != np.double:
                raise ValueError(f"sample_data[{i}] type {sample_data[i].dtype} is not double")
            if sample_indices[i].dtype != np.int32:
                raise ValueError(f"sample_indices[{i}] type {sample_indices[i].dtype} is not int32")

        # c type: double**
        # each double* element points to start of each column of sample data.
1698
        sample_col_ptr: _ctypes_float_array = (ctypes.POINTER(ctypes.c_double) * ncol)()
1699
1700
        # c type int**
        # each int* points to start of indices for each column
1701
        indices_col_ptr: _ctypes_int_array = (ctypes.POINTER(ctypes.c_int32) * ncol)()
1702
        for i in range(ncol):
1703
1704
            sample_col_ptr[i] = _c_float_array(sample_data[i])[0]
            indices_col_ptr[i] = _c_int_array(sample_indices[i])[0]
1705
1706

        num_per_col = np.array([len(d) for d in sample_indices], dtype=np.int32)
1707
        num_per_col_ptr, _, _ = _c_int_array(num_per_col)
1708

1709
        self._handle = ctypes.c_void_p()
1710
        params_str = _param_dict_to_str(self.get_params())
1711
1712
1713
1714
1715
1716
1717
        _safe_call(_LIB.LGBM_DatasetCreateFromSampledColumn(
            ctypes.cast(sample_col_ptr, ctypes.POINTER(ctypes.POINTER(ctypes.c_double))),
            ctypes.cast(indices_col_ptr, ctypes.POINTER(ctypes.POINTER(ctypes.c_int32))),
            ctypes.c_int32(ncol),
            num_per_col_ptr,
            ctypes.c_int32(sample_cnt),
            ctypes.c_int32(total_nrow),
1718
            ctypes.c_int64(total_nrow),
1719
            _c_str(params_str),
1720
            ctypes.byref(self._handle),
1721
1722
1723
1724
1725
1726
1727
1728
1729
1730
1731
1732
1733
1734
1735
1736
1737
1738
        ))
        return self

    def _push_rows(self, data: np.ndarray) -> 'Dataset':
        """Add rows to Dataset.

        Parameters
        ----------
        data : numpy 1-D array
            New data to add to the Dataset.

        Returns
        -------
        self : Dataset
            Dataset object.
        """
        nrow, ncol = data.shape
        data = data.reshape(data.size)
1739
        data_ptr, data_type, _ = _c_float_array(data)
1740
1741

        _safe_call(_LIB.LGBM_DatasetPushRows(
1742
            self._handle,
1743
1744
1745
1746
1747
1748
1749
1750
1751
            data_ptr,
            data_type,
            ctypes.c_int32(nrow),
            ctypes.c_int32(ncol),
            ctypes.c_int32(self._start_row),
        ))
        self._start_row += nrow
        return self

1752
    def get_params(self) -> Dict[str, Any]:
1753
1754
1755
1756
        """Get the used parameters in the Dataset.

        Returns
        -------
1757
        params : dict
1758
1759
1760
1761
1762
1763
1764
1765
1766
1767
1768
1769
1770
1771
1772
            The used parameters in this Dataset object.
        """
        if self.params is not None:
            # no min_data, nthreads and verbose in this function
            dataset_params = _ConfigAliases.get("bin_construct_sample_cnt",
                                                "categorical_feature",
                                                "data_random_seed",
                                                "enable_bundle",
                                                "feature_pre_filter",
                                                "forcedbins_filename",
                                                "group_column",
                                                "header",
                                                "ignore_column",
                                                "is_enable_sparse",
                                                "label_column",
1773
                                                "linear_tree",
1774
1775
1776
1777
                                                "max_bin",
                                                "max_bin_by_feature",
                                                "min_data_in_bin",
                                                "pre_partition",
Nikita Titov's avatar
Nikita Titov committed
1778
                                                "precise_float_parser",
1779
1780
1781
1782
1783
                                                "two_round",
                                                "use_missing",
                                                "weight_column",
                                                "zero_as_missing")
            return {k: v for k, v in self.params.items() if k in dataset_params}
1784
1785
        else:
            return {}
1786

1787
    def _free_handle(self) -> "Dataset":
1788
1789
1790
        if self._handle is not None:
            _safe_call(_LIB.LGBM_DatasetFree(self._handle))
            self._handle = None
1791
        self._need_slice = True
Guolin Ke's avatar
Guolin Ke committed
1792
1793
        if self.used_indices is not None:
            self.data = None
Nikita Titov's avatar
Nikita Titov committed
1794
        return self
wxchan's avatar
wxchan committed
1795

1796
1797
1798
    def _set_init_score_by_predictor(
        self,
        predictor: Optional[_InnerPredictor],
1799
        data: _LGBM_TrainDataType,
1800
        used_indices: Optional[Union[List[int], np.ndarray]]
1801
    ) -> "Dataset":
Guolin Ke's avatar
Guolin Ke committed
1802
        data_has_header = False
1803
        if isinstance(data, (str, Path)) and self.params is not None:
Guolin Ke's avatar
Guolin Ke committed
1804
            # check data has header or not
1805
            data_has_header = any(self.params.get(alias, False) for alias in _ConfigAliases.get("header"))
Guolin Ke's avatar
Guolin Ke committed
1806
        num_data = self.num_data()
1807
        if predictor is not None:
1808
1809
1810
1811
1812
            init_score: Union[np.ndarray, scipy.sparse.spmatrix] = predictor.predict(
                data=data,
                raw_score=True,
                data_has_header=data_has_header
            )
1813
            init_score = init_score.ravel()
1814
            if used_indices is not None:
1815
                assert not self._need_slice
1816
                if isinstance(data, (str, Path)):
1817
                    sub_init_score = np.empty(num_data * predictor.num_class, dtype=np.float64)
1818
                    assert num_data == len(used_indices)
1819
1820
                    for i in range(len(used_indices)):
                        for j in range(predictor.num_class):
1821
1822
1823
1824
                            sub_init_score[i * predictor.num_class + j] = init_score[used_indices[i] * predictor.num_class + j]
                    init_score = sub_init_score
            if predictor.num_class > 1:
                # need to regroup init_score
1825
                new_init_score = np.empty(init_score.size, dtype=np.float64)
1826
1827
                for i in range(num_data):
                    for j in range(predictor.num_class):
1828
1829
1830
                        new_init_score[j * num_data + i] = init_score[i * predictor.num_class + j]
                init_score = new_init_score
        elif self.init_score is not None:
1831
            init_score = np.full_like(self.init_score, fill_value=0.0, dtype=np.float64)
1832
1833
        else:
            return self
Guolin Ke's avatar
Guolin Ke committed
1834
        self.set_init_score(init_score)
1835
        return self
Guolin Ke's avatar
Guolin Ke committed
1836

1837
1838
    def _lazy_init(
        self,
1839
        data: Optional[_LGBM_TrainDataType],
1840
1841
1842
1843
1844
1845
1846
1847
        label: Optional[_LGBM_LabelType],
        reference: Optional["Dataset"],
        weight: Optional[_LGBM_WeightType],
        group: Optional[_LGBM_GroupType],
        init_score: Optional[_LGBM_InitScoreType],
        predictor: Optional[_InnerPredictor],
        feature_name: _LGBM_FeatureNameConfiguration,
        categorical_feature: _LGBM_CategoricalFeatureConfiguration,
1848
1849
        params: Optional[Dict[str, Any]],
        position: Optional[_LGBM_PositionType]
1850
    ) -> "Dataset":
wxchan's avatar
wxchan committed
1851
        if data is None:
1852
            self._handle = None
Nikita Titov's avatar
Nikita Titov committed
1853
            return self
Guolin Ke's avatar
Guolin Ke committed
1854
1855
1856
        if reference is not None:
            self.pandas_categorical = reference.pandas_categorical
            categorical_feature = reference.categorical_feature
1857
1858
1859
1860
        data, feature_name, categorical_feature, self.pandas_categorical = _data_from_pandas(data=data,
                                                                                             feature_name=feature_name,
                                                                                             categorical_feature=categorical_feature,
                                                                                             pandas_categorical=self.pandas_categorical)
Guolin Ke's avatar
Guolin Ke committed
1861

1862
        # process for args
wxchan's avatar
wxchan committed
1863
        params = {} if params is None else params
1864
        args_names = inspect.signature(self.__class__._lazy_init).parameters.keys()
1865
        for key in params.keys():
1866
            if key in args_names:
1867
1868
                _log_warning(f'{key} keyword has been found in `params` and will be ignored.\n'
                             f'Please use {key} argument of the Dataset constructor to pass this parameter.')
1869
        # get categorical features
1870
1871
1872
1873
1874
1875
        if categorical_feature is not None:
            categorical_indices = set()
            feature_dict = {}
            if feature_name is not None:
                feature_dict = {name: i for i, name in enumerate(feature_name)}
            for name in categorical_feature:
1876
                if isinstance(name, str) and name in feature_dict:
1877
                    categorical_indices.add(feature_dict[name])
1878
                elif isinstance(name, int):
1879
1880
                    categorical_indices.add(name)
                else:
1881
                    raise TypeError(f"Wrong type({type(name).__name__}) or unknown name({name}) in categorical_feature")
1882
            if categorical_indices:
1883
1884
                for cat_alias in _ConfigAliases.get("categorical_feature"):
                    if cat_alias in params:
1885
                        # If the params[cat_alias] is equal to categorical_indices, do not report the warning.
1886
                        if not (isinstance(params[cat_alias], list) and set(params[cat_alias]) == categorical_indices):
1887
                            _log_warning(f'{cat_alias} in param dict is overridden.')
1888
                        params.pop(cat_alias, None)
1889
                params['categorical_column'] = sorted(categorical_indices)
1890

1891
        params_str = _param_dict_to_str(params)
1892
        self.params = params
1893
        # process for reference dataset
wxchan's avatar
wxchan committed
1894
        ref_dataset = None
wxchan's avatar
wxchan committed
1895
        if isinstance(reference, Dataset):
1896
            ref_dataset = reference.construct()._handle
wxchan's avatar
wxchan committed
1897
1898
        elif reference is not None:
            raise TypeError('Reference dataset should be None or dataset instance')
1899
        # start construct data
1900
        if isinstance(data, (str, Path)):
1901
            self._handle = ctypes.c_void_p()
wxchan's avatar
wxchan committed
1902
            _safe_call(_LIB.LGBM_DatasetCreateFromFile(
1903
1904
                _c_str(str(data)),
                _c_str(params_str),
wxchan's avatar
wxchan committed
1905
                ref_dataset,
1906
                ctypes.byref(self._handle)))
wxchan's avatar
wxchan committed
1907
1908
        elif isinstance(data, scipy.sparse.csr_matrix):
            self.__init_from_csr(data, params_str, ref_dataset)
Guolin Ke's avatar
Guolin Ke committed
1909
1910
        elif isinstance(data, scipy.sparse.csc_matrix):
            self.__init_from_csc(data, params_str, ref_dataset)
wxchan's avatar
wxchan committed
1911
1912
        elif isinstance(data, np.ndarray):
            self.__init_from_np2d(data, params_str, ref_dataset)
1913
1914
1915
1916
1917
1918
1919
1920
1921
        elif isinstance(data, list) and len(data) > 0:
            if all(isinstance(x, np.ndarray) for x in data):
                self.__init_from_list_np2d(data, params_str, ref_dataset)
            elif all(isinstance(x, Sequence) for x in data):
                self.__init_from_seqs(data, ref_dataset)
            else:
                raise TypeError('Data list can only be of ndarray or Sequence')
        elif isinstance(data, Sequence):
            self.__init_from_seqs([data], ref_dataset)
1922
        elif isinstance(data, dt_DataTable):
1923
            self.__init_from_np2d(data.to_numpy(), params_str, ref_dataset)
wxchan's avatar
wxchan committed
1924
1925
1926
1927
        else:
            try:
                csr = scipy.sparse.csr_matrix(data)
                self.__init_from_csr(csr, params_str, ref_dataset)
1928
1929
            except BaseException as err:
                raise TypeError(f'Cannot initialize Dataset from {type(data).__name__}') from err
wxchan's avatar
wxchan committed
1930
1931
1932
        if label is not None:
            self.set_label(label)
        if self.get_label() is None:
1933
            raise ValueError("Label should not be None")
wxchan's avatar
wxchan committed
1934
1935
1936
1937
        if weight is not None:
            self.set_weight(weight)
        if group is not None:
            self.set_group(group)
1938
1939
        if position is not None:
            self.set_position(position)
1940
1941
        if isinstance(predictor, _InnerPredictor):
            if self._predictor is None and init_score is not None:
1942
                _log_warning("The init_score will be overridden by the prediction of init_model.")
1943
1944
1945
1946
1947
            self._set_init_score_by_predictor(
                predictor=predictor,
                data=data,
                used_indices=None
            )
1948
1949
        elif init_score is not None:
            self.set_init_score(init_score)
Guolin Ke's avatar
Guolin Ke committed
1950
        elif predictor is not None:
1951
            raise TypeError(f'Wrong predictor type {type(predictor).__name__}')
Guolin Ke's avatar
Guolin Ke committed
1952
        # set feature names
Nikita Titov's avatar
Nikita Titov committed
1953
        return self.set_feature_name(feature_name)
wxchan's avatar
wxchan committed
1954

1955
1956
    @staticmethod
    def _yield_row_from_seqlist(seqs: List[Sequence], indices: Iterable[int]):
1957
1958
1959
1960
1961
1962
1963
1964
1965
1966
1967
1968
1969
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
1980
1981
        offset = 0
        seq_id = 0
        seq = seqs[seq_id]
        for row_id in indices:
            assert row_id >= offset, "sample indices are expected to be monotonic"
            while row_id >= offset + len(seq):
                offset += len(seq)
                seq_id += 1
                seq = seqs[seq_id]
            id_in_seq = row_id - offset
            row = seq[id_in_seq]
            yield row if row.flags['OWNDATA'] else row.copy()

    def __sample(self, seqs: List[Sequence], total_nrow: int) -> Tuple[List[np.ndarray], List[np.ndarray]]:
        """Sample data from seqs.

        Mimics behavior in c_api.cpp:LGBM_DatasetCreateFromMats()

        Returns
        -------
            sampled_rows, sampled_row_indices
        """
        indices = self._create_sample_indices(total_nrow)

        # Select sampled rows, transpose to column order.
1982
        sampled = np.array(list(self._yield_row_from_seqlist(seqs, indices)))
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
        sampled = sampled.T

        filtered = []
        filtered_idx = []
        sampled_row_range = np.arange(len(indices), dtype=np.int32)
        for col in sampled:
            col_predicate = (np.abs(col) > ZERO_THRESHOLD) | np.isnan(col)
            filtered_col = col[col_predicate]
            filtered_row_idx = sampled_row_range[col_predicate]

            filtered.append(filtered_col)
            filtered_idx.append(filtered_row_idx)

        return filtered, filtered_idx

1998
1999
2000
    def __init_from_seqs(
        self,
        seqs: List[Sequence],
2001
        ref_dataset: Optional[_DatasetHandle]
2002
    ) -> "Dataset":
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
        """
        Initialize data from list of Sequence objects.

        Sequence: Generic Data Access Object
            Supports random access and access by batch if properly defined by user

        Data scheme uniformity are trusted, not checked
        """
        total_nrow = sum(len(seq) for seq in seqs)

        # create validation dataset from ref_dataset
        if ref_dataset is not None:
            self._init_from_ref_dataset(total_nrow, ref_dataset)
        else:
2017
            param_str = _param_dict_to_str(self.get_params())
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
            sample_cnt = _get_sample_count(total_nrow, param_str)

            sample_data, col_indices = self.__sample(seqs, total_nrow)
            self._init_from_sample(sample_data, col_indices, sample_cnt, total_nrow)

        for seq in seqs:
            nrow = len(seq)
            batch_size = getattr(seq, 'batch_size', None) or Sequence.batch_size
            for start in range(0, nrow, batch_size):
                end = min(start + batch_size, nrow)
                self._push_rows(seq[start:end])
        return self

2031
2032
2033
2034
2035
2036
    def __init_from_np2d(
        self,
        mat: np.ndarray,
        params_str: str,
        ref_dataset: Optional[_DatasetHandle]
    ) -> "Dataset":
2037
        """Initialize data from a 2-D numpy matrix."""
wxchan's avatar
wxchan committed
2038
2039
2040
        if len(mat.shape) != 2:
            raise ValueError('Input numpy.ndarray must be 2 dimensional')

2041
        self._handle = ctypes.c_void_p()
wxchan's avatar
wxchan committed
2042
2043
        if mat.dtype == np.float32 or mat.dtype == np.float64:
            data = np.array(mat.reshape(mat.size), dtype=mat.dtype, copy=False)
2044
        else:  # change non-float data to float data, need to copy
wxchan's avatar
wxchan committed
2045
2046
            data = np.array(mat.reshape(mat.size), dtype=np.float32)

2047
        ptr_data, type_ptr_data, _ = _c_float_array(data)
wxchan's avatar
wxchan committed
2048
2049
        _safe_call(_LIB.LGBM_DatasetCreateFromMat(
            ptr_data,
Guolin Ke's avatar
Guolin Ke committed
2050
            ctypes.c_int(type_ptr_data),
2051
2052
            ctypes.c_int32(mat.shape[0]),
            ctypes.c_int32(mat.shape[1]),
2053
            ctypes.c_int(_C_API_IS_ROW_MAJOR),
2054
            _c_str(params_str),
wxchan's avatar
wxchan committed
2055
            ref_dataset,
2056
            ctypes.byref(self._handle)))
Nikita Titov's avatar
Nikita Titov committed
2057
        return self
wxchan's avatar
wxchan committed
2058

2059
2060
2061
2062
2063
2064
    def __init_from_list_np2d(
        self,
        mats: List[np.ndarray],
        params_str: str,
        ref_dataset: Optional[_DatasetHandle]
    ) -> "Dataset":
2065
        """Initialize data from a list of 2-D numpy matrices."""
2066
        ncol = mats[0].shape[1]
2067
        nrow = np.empty((len(mats),), np.int32)
2068
        ptr_data: _ctypes_float_array
2069
2070
2071
2072
2073
2074
        if mats[0].dtype == np.float64:
            ptr_data = (ctypes.POINTER(ctypes.c_double) * len(mats))()
        else:
            ptr_data = (ctypes.POINTER(ctypes.c_float) * len(mats))()

        holders = []
2075
        type_ptr_data = -1
2076
2077
2078
2079
2080
2081
2082
2083
2084
2085
2086
2087

        for i, mat in enumerate(mats):
            if len(mat.shape) != 2:
                raise ValueError('Input numpy.ndarray must be 2 dimensional')

            if mat.shape[1] != ncol:
                raise ValueError('Input arrays must have same number of columns')

            nrow[i] = mat.shape[0]

            if mat.dtype == np.float32 or mat.dtype == np.float64:
                mats[i] = np.array(mat.reshape(mat.size), dtype=mat.dtype, copy=False)
2088
            else:  # change non-float data to float data, need to copy
2089
2090
                mats[i] = np.array(mat.reshape(mat.size), dtype=np.float32)

2091
            chunk_ptr_data, chunk_type_ptr_data, holder = _c_float_array(mats[i])
2092
            if type_ptr_data != -1 and chunk_type_ptr_data != type_ptr_data:
2093
2094
2095
2096
2097
                raise ValueError('Input chunks must have same type')
            ptr_data[i] = chunk_ptr_data
            type_ptr_data = chunk_type_ptr_data
            holders.append(holder)

2098
        self._handle = ctypes.c_void_p()
2099
        _safe_call(_LIB.LGBM_DatasetCreateFromMats(
2100
            ctypes.c_int32(len(mats)),
2101
2102
2103
            ctypes.cast(ptr_data, ctypes.POINTER(ctypes.POINTER(ctypes.c_double))),
            ctypes.c_int(type_ptr_data),
            nrow.ctypes.data_as(ctypes.POINTER(ctypes.c_int32)),
2104
            ctypes.c_int32(ncol),
2105
            ctypes.c_int(_C_API_IS_ROW_MAJOR),
2106
            _c_str(params_str),
2107
            ref_dataset,
2108
            ctypes.byref(self._handle)))
Nikita Titov's avatar
Nikita Titov committed
2109
        return self
2110

2111
2112
2113
2114
2115
2116
    def __init_from_csr(
        self,
        csr: scipy.sparse.csr_matrix,
        params_str: str,
        ref_dataset: Optional[_DatasetHandle]
    ) -> "Dataset":
2117
        """Initialize data from a CSR matrix."""
wxchan's avatar
wxchan committed
2118
        if len(csr.indices) != len(csr.data):
2119
            raise ValueError(f'Length mismatch: {len(csr.indices)} vs {len(csr.data)}')
2120
        self._handle = ctypes.c_void_p()
wxchan's avatar
wxchan committed
2121

2122
2123
        ptr_indptr, type_ptr_indptr, __ = _c_int_array(csr.indptr)
        ptr_data, type_ptr_data, _ = _c_float_array(csr.data)
wxchan's avatar
wxchan committed
2124

2125
        assert csr.shape[1] <= _MAX_INT32
2126
        csr_indices = csr.indices.astype(np.int32, copy=False)
2127

wxchan's avatar
wxchan committed
2128
2129
        _safe_call(_LIB.LGBM_DatasetCreateFromCSR(
            ptr_indptr,
Guolin Ke's avatar
Guolin Ke committed
2130
            ctypes.c_int(type_ptr_indptr),
2131
            csr_indices.ctypes.data_as(ctypes.POINTER(ctypes.c_int32)),
wxchan's avatar
wxchan committed
2132
            ptr_data,
Guolin Ke's avatar
Guolin Ke committed
2133
2134
2135
2136
            ctypes.c_int(type_ptr_data),
            ctypes.c_int64(len(csr.indptr)),
            ctypes.c_int64(len(csr.data)),
            ctypes.c_int64(csr.shape[1]),
2137
            _c_str(params_str),
wxchan's avatar
wxchan committed
2138
            ref_dataset,
2139
            ctypes.byref(self._handle)))
Nikita Titov's avatar
Nikita Titov committed
2140
        return self
wxchan's avatar
wxchan committed
2141

2142
2143
2144
2145
2146
2147
    def __init_from_csc(
        self,
        csc: scipy.sparse.csc_matrix,
        params_str: str,
        ref_dataset: Optional[_DatasetHandle]
    ) -> "Dataset":
2148
        """Initialize data from a CSC matrix."""
Guolin Ke's avatar
Guolin Ke committed
2149
        if len(csc.indices) != len(csc.data):
2150
            raise ValueError(f'Length mismatch: {len(csc.indices)} vs {len(csc.data)}')
2151
        self._handle = ctypes.c_void_p()
Guolin Ke's avatar
Guolin Ke committed
2152

2153
2154
        ptr_indptr, type_ptr_indptr, __ = _c_int_array(csc.indptr)
        ptr_data, type_ptr_data, _ = _c_float_array(csc.data)
Guolin Ke's avatar
Guolin Ke committed
2155

2156
        assert csc.shape[0] <= _MAX_INT32
2157
        csc_indices = csc.indices.astype(np.int32, copy=False)
2158

Guolin Ke's avatar
Guolin Ke committed
2159
2160
        _safe_call(_LIB.LGBM_DatasetCreateFromCSC(
            ptr_indptr,
Guolin Ke's avatar
Guolin Ke committed
2161
            ctypes.c_int(type_ptr_indptr),
2162
            csc_indices.ctypes.data_as(ctypes.POINTER(ctypes.c_int32)),
Guolin Ke's avatar
Guolin Ke committed
2163
            ptr_data,
Guolin Ke's avatar
Guolin Ke committed
2164
2165
2166
2167
            ctypes.c_int(type_ptr_data),
            ctypes.c_int64(len(csc.indptr)),
            ctypes.c_int64(len(csc.data)),
            ctypes.c_int64(csc.shape[0]),
2168
            _c_str(params_str),
Guolin Ke's avatar
Guolin Ke committed
2169
            ref_dataset,
2170
            ctypes.byref(self._handle)))
Nikita Titov's avatar
Nikita Titov committed
2171
        return self
Guolin Ke's avatar
Guolin Ke committed
2172

2173
    @staticmethod
2174
2175
2176
2177
2178
2179
    def _compare_params_for_warning(
        params: Optional[Dict[str, Any]],
        other_params: Optional[Dict[str, Any]],
        ignore_keys: Set[str]
    ) -> bool:
        """Compare two dictionaries with params ignoring some keys.
2180

2181
2182
2183
2184
2185
2186
2187
2188
2189
2190
        It is only for the warning purpose.

        Parameters
        ----------
        params : dict or None
            One dictionary with parameters to compare.
        other_params : dict or None
            Another dictionary with parameters to compare.
        ignore_keys : set
            Keys that should be ignored during comparing two dictionaries.
2191
2192
2193

        Returns
        -------
2194
2195
        compare_result : bool
          Returns whether two dictionaries with params are equal.
2196
2197
2198
2199
2200
2201
2202
2203
2204
2205
2206
2207
2208
2209
2210
        """
        if params is None:
            params = {}
        if other_params is None:
            other_params = {}
        for k in other_params:
            if k not in ignore_keys:
                if k not in params or params[k] != other_params[k]:
                    return False
        for k in params:
            if k not in ignore_keys:
                if k not in other_params or params[k] != other_params[k]:
                    return False
        return True

2211
    def construct(self) -> "Dataset":
2212
2213
2214
2215
2216
        """Lazy init.

        Returns
        -------
        self : Dataset
Nikita Titov's avatar
Nikita Titov committed
2217
            Constructed Dataset object.
2218
        """
2219
        if self._handle is None:
wxchan's avatar
wxchan committed
2220
            if self.reference is not None:
2221
                reference_params = self.reference.get_params()
2222
2223
                params = self.get_params()
                if params != reference_params:
2224
2225
2226
2227
2228
                    if not self._compare_params_for_warning(
                        params=params,
                        other_params=reference_params,
                        ignore_keys=_ConfigAliases.get("categorical_feature")
                    ):
2229
                        _log_warning('Overriding the parameters from Reference Dataset.')
2230
                    self._update_params(reference_params)
wxchan's avatar
wxchan committed
2231
                if self.used_indices is None:
2232
                    # create valid
2233
                    self._lazy_init(data=self.data, label=self.label, reference=self.reference,
2234
                                    weight=self.weight, group=self.group, position=self.position,
2235
                                    init_score=self.init_score, predictor=self._predictor,
2236
                                    feature_name=self.feature_name, categorical_feature='auto', params=self.params)
wxchan's avatar
wxchan committed
2237
                else:
2238
                    # construct subset
2239
                    used_indices = _list_to_1d_numpy(self.used_indices, dtype=np.int32, name='used_indices')
2240
                    assert used_indices.flags.c_contiguous
Guolin Ke's avatar
Guolin Ke committed
2241
                    if self.reference.group is not None:
2242
                        group_info = np.array(self.reference.group).astype(np.int32, copy=False)
2243
                        _, self.group = np.unique(np.repeat(range(len(group_info)), repeats=group_info)[self.used_indices],
2244
                                                  return_counts=True)
2245
                    self._handle = ctypes.c_void_p()
2246
                    params_str = _param_dict_to_str(self.params)
wxchan's avatar
wxchan committed
2247
                    _safe_call(_LIB.LGBM_DatasetGetSubset(
2248
                        self.reference.construct()._handle,
wxchan's avatar
wxchan committed
2249
                        used_indices.ctypes.data_as(ctypes.POINTER(ctypes.c_int32)),
2250
                        ctypes.c_int32(used_indices.shape[0]),
2251
                        _c_str(params_str),
2252
                        ctypes.byref(self._handle)))
Guolin Ke's avatar
Guolin Ke committed
2253
2254
                    if not self.free_raw_data:
                        self.get_data()
Guolin Ke's avatar
Guolin Ke committed
2255
2256
                    if self.group is not None:
                        self.set_group(self.group)
2257
2258
                    if self.position is not None:
                        self.set_position(self.position)
wxchan's avatar
wxchan committed
2259
2260
                    if self.get_label() is None:
                        raise ValueError("Label should not be None.")
Guolin Ke's avatar
Guolin Ke committed
2261
2262
                    if isinstance(self._predictor, _InnerPredictor) and self._predictor is not self.reference._predictor:
                        self.get_data()
2263
2264
2265
2266
2267
                        self._set_init_score_by_predictor(
                            predictor=self._predictor,
                            data=self.data,
                            used_indices=used_indices
                        )
wxchan's avatar
wxchan committed
2268
            else:
2269
                # create train
2270
                self._lazy_init(data=self.data, label=self.label, reference=None,
2271
2272
                                weight=self.weight, group=self.group,
                                init_score=self.init_score, predictor=self._predictor,
2273
2274
                                feature_name=self.feature_name, categorical_feature=self.categorical_feature,
                                params=self.params, position=self.position)
wxchan's avatar
wxchan committed
2275
2276
            if self.free_raw_data:
                self.data = None
2277
            self.feature_name = self.get_feature_name()
wxchan's avatar
wxchan committed
2278
        return self
wxchan's avatar
wxchan committed
2279

2280
2281
    def create_valid(
        self,
2282
        data: _LGBM_TrainDataType,
2283
        label: Optional[_LGBM_LabelType] = None,
2284
2285
2286
        weight: Optional[_LGBM_WeightType] = None,
        group: Optional[_LGBM_GroupType] = None,
        init_score: Optional[_LGBM_InitScoreType] = None,
2287
2288
        params: Optional[Dict[str, Any]] = None,
        position: Optional[_LGBM_PositionType] = None
2289
    ) -> "Dataset":
2290
        """Create validation data align with current Dataset.
wxchan's avatar
wxchan committed
2291
2292
2293

        Parameters
        ----------
2294
        data : str, pathlib.Path, numpy array, pandas DataFrame, H2O DataTable's Frame, scipy.sparse, Sequence, list of Sequence or list of numpy array
wxchan's avatar
wxchan committed
2295
            Data source of Dataset.
2296
            If str or pathlib.Path, it represents the path to a text file (CSV, TSV, or LibSVM) or a LightGBM Dataset binary file.
2297
        label : list, numpy 1-D array, pandas Series / one-column DataFrame or None, optional (default=None)
2298
2299
            Label of the data.
        weight : list, numpy 1-D array, pandas Series or None, optional (default=None)
2300
            Weight for each instance. Weights should be non-negative.
2301
        group : list, numpy 1-D array, pandas Series or None, optional (default=None)
2302
2303
2304
            Group/query data.
            Only used in the learning-to-rank task.
            sum(group) = n_samples.
2305
2306
            For example, if you have a 100-document dataset with ``group = [10, 20, 40, 10, 10, 10]``, that means that you have 6 groups,
            where the first 10 records are in the first group, records 11-30 are in the second group, records 31-70 are in the third group, etc.
2307
        init_score : list, list of lists (for multi-class task), numpy array, pandas Series, pandas DataFrame (for multi-class task), or None, optional (default=None)
2308
            Init score for Dataset.
Nikita Titov's avatar
Nikita Titov committed
2309
        params : dict or None, optional (default=None)
2310
            Other parameters for validation Dataset.
2311
2312
        position : numpy 1-D array, pandas Series or None, optional (default=None)
            Position of items used in unbiased learning-to-rank task.
2313
2314
2315

        Returns
        -------
Nikita Titov's avatar
Nikita Titov committed
2316
2317
        valid : Dataset
            Validation Dataset with reference to self.
wxchan's avatar
wxchan committed
2318
        """
2319
        ret = Dataset(data, label=label, reference=self,
2320
                      weight=weight, group=group, position=position, init_score=init_score,
2321
                      params=params, free_raw_data=self.free_raw_data)
wxchan's avatar
wxchan committed
2322
        ret._predictor = self._predictor
2323
        ret.pandas_categorical = self.pandas_categorical
wxchan's avatar
wxchan committed
2324
        return ret
wxchan's avatar
wxchan committed
2325

2326
2327
2328
2329
2330
    def subset(
        self,
        used_indices: List[int],
        params: Optional[Dict[str, Any]] = None
    ) -> "Dataset":
2331
        """Get subset of current Dataset.
wxchan's avatar
wxchan committed
2332
2333
2334
2335

        Parameters
        ----------
        used_indices : list of int
2336
            Indices used to create the subset.
Nikita Titov's avatar
Nikita Titov committed
2337
        params : dict or None, optional (default=None)
2338
            These parameters will be passed to Dataset constructor.
2339
2340
2341
2342
2343

        Returns
        -------
        subset : Dataset
            Subset of the current Dataset.
wxchan's avatar
wxchan committed
2344
        """
wxchan's avatar
wxchan committed
2345
2346
        if params is None:
            params = self.params
wxchan's avatar
wxchan committed
2347
        ret = Dataset(None, reference=self, feature_name=self.feature_name,
2348
2349
                      categorical_feature=self.categorical_feature, params=params,
                      free_raw_data=self.free_raw_data)
wxchan's avatar
wxchan committed
2350
        ret._predictor = self._predictor
2351
        ret.pandas_categorical = self.pandas_categorical
2352
        ret.used_indices = sorted(used_indices)
wxchan's avatar
wxchan committed
2353
2354
        return ret

2355
    def save_binary(self, filename: Union[str, Path]) -> "Dataset":
2356
        """Save Dataset to a binary file.
wxchan's avatar
wxchan committed
2357

2358
2359
2360
2361
2362
        .. note::

            Please note that `init_score` is not saved in binary file.
            If you need it, please set it again after loading Dataset.

wxchan's avatar
wxchan committed
2363
2364
        Parameters
        ----------
2365
        filename : str or pathlib.Path
wxchan's avatar
wxchan committed
2366
            Name of the output file.
Nikita Titov's avatar
Nikita Titov committed
2367
2368
2369
2370
2371

        Returns
        -------
        self : Dataset
            Returns self.
wxchan's avatar
wxchan committed
2372
2373
        """
        _safe_call(_LIB.LGBM_DatasetSaveBinary(
2374
            self.construct()._handle,
2375
            _c_str(str(filename))))
Nikita Titov's avatar
Nikita Titov committed
2376
        return self
wxchan's avatar
wxchan committed
2377

2378
    def _update_params(self, params: Optional[Dict[str, Any]]) -> "Dataset":
2379
2380
        if not params:
            return self
2381
        params = deepcopy(params)
2382
2383
2384
2385
2386

        def update():
            if not self.params:
                self.params = params
            else:
2387
                self._params_back_up = deepcopy(self.params)
2388
2389
                self.params.update(params)

2390
        if self._handle is None:
2391
2392
2393
            update()
        elif params is not None:
            ret = _LIB.LGBM_DatasetUpdateParamChecking(
2394
2395
                _c_str(_param_dict_to_str(self.params)),
                _c_str(_param_dict_to_str(params)))
2396
2397
2398
2399
2400
2401
            if ret != 0:
                # could be updated if data is not freed
                if self.data is not None:
                    update()
                    self._free_handle()
                else:
2402
                    raise LightGBMError(_LIB.LGBM_GetLastError().decode('utf-8'))
Nikita Titov's avatar
Nikita Titov committed
2403
        return self
wxchan's avatar
wxchan committed
2404

2405
    def _reverse_update_params(self) -> "Dataset":
2406
        if self._handle is None:
2407
2408
            self.params = deepcopy(self._params_back_up)
            self._params_back_up = None
Nikita Titov's avatar
Nikita Titov committed
2409
        return self
2410

2411
2412
2413
    def set_field(
        self,
        field_name: str,
2414
        data: Optional[Union[List[List[float]], List[List[int]], List[float], List[int], np.ndarray, pd_Series, pd_DataFrame]]
2415
    ) -> "Dataset":
wxchan's avatar
wxchan committed
2416
        """Set property into the Dataset.
wxchan's avatar
wxchan committed
2417
2418
2419

        Parameters
        ----------
2420
        field_name : str
2421
            The field name of the information.
2422
2423
        data : list, list of lists (for multi-class task), numpy array, pandas Series, pandas DataFrame (for multi-class task), or None
            The data to be set.
Nikita Titov's avatar
Nikita Titov committed
2424
2425
2426
2427
2428

        Returns
        -------
        self : Dataset
            Dataset with set property.
wxchan's avatar
wxchan committed
2429
        """
2430
        if self._handle is None:
2431
            raise Exception(f"Cannot set {field_name} before construct dataset")
wxchan's avatar
wxchan committed
2432
        if data is None:
2433
            # set to None
wxchan's avatar
wxchan committed
2434
            _safe_call(_LIB.LGBM_DatasetSetField(
2435
                self._handle,
2436
                _c_str(field_name),
wxchan's avatar
wxchan committed
2437
                None,
Guolin Ke's avatar
Guolin Ke committed
2438
                ctypes.c_int(0),
2439
                ctypes.c_int(_FIELD_TYPE_MAPPER[field_name])))
Nikita Titov's avatar
Nikita Titov committed
2440
            return self
2441
        dtype: "np.typing.DTypeLike"
2442
        if field_name == 'init_score':
Guolin Ke's avatar
Guolin Ke committed
2443
            dtype = np.float64
2444
            if _is_1d_collection(data):
2445
                data = _list_to_1d_numpy(data, dtype=dtype, name=field_name)
2446
            elif _is_2d_collection(data):
2447
                data = _data_to_2d_numpy(data, dtype=dtype, name=field_name)
2448
2449
2450
2451
2452
2453
2454
                data = data.ravel(order='F')
            else:
                raise TypeError(
                    'init_score must be list, numpy 1-D array or pandas Series.\n'
                    'In multiclass classification init_score can also be a list of lists, numpy 2-D array or pandas DataFrame.'
                )
        else:
2455
            dtype = np.int32 if (field_name == 'group' or field_name == 'position') else np.float32
2456
            data = _list_to_1d_numpy(data, dtype=dtype, name=field_name)
2457

2458
        ptr_data: Union[_ctypes_float_ptr, _ctypes_int_ptr]
2459
        if data.dtype == np.float32 or data.dtype == np.float64:
2460
            ptr_data, type_data, _ = _c_float_array(data)
wxchan's avatar
wxchan committed
2461
        elif data.dtype == np.int32:
2462
            ptr_data, type_data, _ = _c_int_array(data)
wxchan's avatar
wxchan committed
2463
        else:
2464
            raise TypeError(f"Expected np.float32/64 or np.int32, met type({data.dtype})")
2465
        if type_data != _FIELD_TYPE_MAPPER[field_name]:
2466
            raise TypeError("Input type error for set_field")
wxchan's avatar
wxchan committed
2467
        _safe_call(_LIB.LGBM_DatasetSetField(
2468
            self._handle,
2469
            _c_str(field_name),
wxchan's avatar
wxchan committed
2470
            ptr_data,
Guolin Ke's avatar
Guolin Ke committed
2471
2472
            ctypes.c_int(len(data)),
            ctypes.c_int(type_data)))
2473
        self.version += 1
Nikita Titov's avatar
Nikita Titov committed
2474
        return self
wxchan's avatar
wxchan committed
2475

2476
    def get_field(self, field_name: str) -> Optional[np.ndarray]:
wxchan's avatar
wxchan committed
2477
        """Get property from the Dataset.
wxchan's avatar
wxchan committed
2478
2479
2480

        Parameters
        ----------
2481
        field_name : str
2482
            The field name of the information.
wxchan's avatar
wxchan committed
2483
2484
2485

        Returns
        -------
2486
        info : numpy array or None
2487
            A numpy array with information from the Dataset.
Guolin Ke's avatar
Guolin Ke committed
2488
        """
2489
        if self._handle is None:
2490
            raise Exception(f"Cannot get {field_name} before construct Dataset")
2491
2492
        tmp_out_len = ctypes.c_int(0)
        out_type = ctypes.c_int(0)
wxchan's avatar
wxchan committed
2493
2494
        ret = ctypes.POINTER(ctypes.c_void_p)()
        _safe_call(_LIB.LGBM_DatasetGetField(
2495
            self._handle,
2496
            _c_str(field_name),
wxchan's avatar
wxchan committed
2497
2498
2499
            ctypes.byref(tmp_out_len),
            ctypes.byref(ret),
            ctypes.byref(out_type)))
2500
        if out_type.value != _FIELD_TYPE_MAPPER[field_name]:
wxchan's avatar
wxchan committed
2501
2502
2503
            raise TypeError("Return type error for get_field")
        if tmp_out_len.value == 0:
            return None
2504
        if out_type.value == _C_API_DTYPE_INT32:
2505
            arr = _cint32_array_to_numpy(ctypes.cast(ret, ctypes.POINTER(ctypes.c_int32)), tmp_out_len.value)
2506
        elif out_type.value == _C_API_DTYPE_FLOAT32:
2507
            arr = _cfloat32_array_to_numpy(ctypes.cast(ret, ctypes.POINTER(ctypes.c_float)), tmp_out_len.value)
2508
        elif out_type.value == _C_API_DTYPE_FLOAT64:
2509
            arr = _cfloat64_array_to_numpy(ctypes.cast(ret, ctypes.POINTER(ctypes.c_double)), tmp_out_len.value)
2510
        else:
wxchan's avatar
wxchan committed
2511
            raise TypeError("Unknown type")
2512
2513
2514
2515
2516
2517
        if field_name == 'init_score':
            num_data = self.num_data()
            num_classes = arr.size // num_data
            if num_classes > 1:
                arr = arr.reshape((num_data, num_classes), order='F')
        return arr
Guolin Ke's avatar
Guolin Ke committed
2518

2519
2520
    def set_categorical_feature(
        self,
2521
        categorical_feature: _LGBM_CategoricalFeatureConfiguration
2522
    ) -> "Dataset":
2523
        """Set categorical features.
2524
2525
2526

        Parameters
        ----------
2527
        categorical_feature : list of str or int, or 'auto'
2528
            Names or indices of categorical features.
Nikita Titov's avatar
Nikita Titov committed
2529
2530
2531
2532
2533

        Returns
        -------
        self : Dataset
            Dataset with set categorical features.
2534
2535
        """
        if self.categorical_feature == categorical_feature:
Nikita Titov's avatar
Nikita Titov committed
2536
            return self
2537
        if self.data is not None:
2538
2539
            if self.categorical_feature is None:
                self.categorical_feature = categorical_feature
Nikita Titov's avatar
Nikita Titov committed
2540
                return self._free_handle()
2541
            elif categorical_feature == 'auto':
Nikita Titov's avatar
Nikita Titov committed
2542
                return self
2543
            else:
2544
2545
                if self.categorical_feature != 'auto':
                    _log_warning('categorical_feature in Dataset is overridden.\n'
2546
                                 f'New categorical_feature is {list(categorical_feature)}')
2547
                self.categorical_feature = categorical_feature
Nikita Titov's avatar
Nikita Titov committed
2548
                return self._free_handle()
2549
        else:
2550
2551
            raise LightGBMError("Cannot set categorical feature after freed raw data, "
                                "set free_raw_data=False when construct Dataset to avoid this.")
2552

2553
2554
2555
2556
    def _set_predictor(
        self,
        predictor: Optional[_InnerPredictor]
    ) -> "Dataset":
2557
2558
2559
2560
        """Set predictor for continued training.

        It is not recommended for user to call this function.
        Please use init_model argument in engine.train() or engine.cv() instead.
Guolin Ke's avatar
Guolin Ke committed
2561
        """
2562
        if predictor is None and self._predictor is None:
Nikita Titov's avatar
Nikita Titov committed
2563
            return self
2564
2565
2566
        elif isinstance(predictor, _InnerPredictor) and isinstance(self._predictor, _InnerPredictor):
            if (predictor == self._predictor) and (predictor.current_iteration() == self._predictor.current_iteration()):
                return self
2567
        if self._handle is None:
Guolin Ke's avatar
Guolin Ke committed
2568
            self._predictor = predictor
2569
2570
        elif self.data is not None:
            self._predictor = predictor
2571
2572
2573
2574
2575
            self._set_init_score_by_predictor(
                predictor=self._predictor,
                data=self.data,
                used_indices=None
            )
2576
2577
        elif self.used_indices is not None and self.reference is not None and self.reference.data is not None:
            self._predictor = predictor
2578
2579
2580
2581
2582
            self._set_init_score_by_predictor(
                predictor=self._predictor,
                data=self.reference.data,
                used_indices=self.used_indices
            )
Guolin Ke's avatar
Guolin Ke committed
2583
        else:
2584
2585
            raise LightGBMError("Cannot set predictor after freed raw data, "
                                "set free_raw_data=False when construct Dataset to avoid this.")
2586
        return self
Guolin Ke's avatar
Guolin Ke committed
2587

2588
    def set_reference(self, reference: "Dataset") -> "Dataset":
2589
        """Set reference Dataset.
Guolin Ke's avatar
Guolin Ke committed
2590
2591
2592
2593

        Parameters
        ----------
        reference : Dataset
2594
            Reference that is used as a template to construct the current Dataset.
Nikita Titov's avatar
Nikita Titov committed
2595
2596
2597
2598
2599

        Returns
        -------
        self : Dataset
            Dataset with set reference.
Guolin Ke's avatar
Guolin Ke committed
2600
        """
2601
2602
2603
        self.set_categorical_feature(reference.categorical_feature) \
            .set_feature_name(reference.feature_name) \
            ._set_predictor(reference._predictor)
2604
        # we're done if self and reference share a common upstream reference
2605
        if self.get_ref_chain().intersection(reference.get_ref_chain()):
Nikita Titov's avatar
Nikita Titov committed
2606
            return self
Guolin Ke's avatar
Guolin Ke committed
2607
2608
        if self.data is not None:
            self.reference = reference
Nikita Titov's avatar
Nikita Titov committed
2609
            return self._free_handle()
Guolin Ke's avatar
Guolin Ke committed
2610
        else:
2611
2612
            raise LightGBMError("Cannot set reference after freed raw data, "
                                "set free_raw_data=False when construct Dataset to avoid this.")
Guolin Ke's avatar
Guolin Ke committed
2613

2614
    def set_feature_name(self, feature_name: _LGBM_FeatureNameConfiguration) -> "Dataset":
2615
        """Set feature name.
Guolin Ke's avatar
Guolin Ke committed
2616
2617
2618

        Parameters
        ----------
2619
        feature_name : list of str
2620
            Feature names.
Nikita Titov's avatar
Nikita Titov committed
2621
2622
2623
2624
2625

        Returns
        -------
        self : Dataset
            Dataset with set feature name.
Guolin Ke's avatar
Guolin Ke committed
2626
        """
2627
2628
        if feature_name != 'auto':
            self.feature_name = feature_name
2629
        if self._handle is not None and feature_name is not None and feature_name != 'auto':
wxchan's avatar
wxchan committed
2630
            if len(feature_name) != self.num_feature():
2631
                raise ValueError(f"Length of feature_name({len(feature_name)}) and num_feature({self.num_feature()}) don't match")
2632
            c_feature_name = [_c_str(name) for name in feature_name]
wxchan's avatar
wxchan committed
2633
            _safe_call(_LIB.LGBM_DatasetSetFeatureNames(
2634
                self._handle,
2635
                _c_array(ctypes.c_char_p, c_feature_name),
Guolin Ke's avatar
Guolin Ke committed
2636
                ctypes.c_int(len(feature_name))))
Nikita Titov's avatar
Nikita Titov committed
2637
        return self
Guolin Ke's avatar
Guolin Ke committed
2638

2639
    def set_label(self, label: Optional[_LGBM_LabelType]) -> "Dataset":
2640
        """Set label of Dataset.
Guolin Ke's avatar
Guolin Ke committed
2641
2642
2643

        Parameters
        ----------
2644
        label : list, numpy 1-D array, pandas Series / one-column DataFrame or None
2645
            The label information to be set into Dataset.
Nikita Titov's avatar
Nikita Titov committed
2646
2647
2648
2649
2650

        Returns
        -------
        self : Dataset
            Dataset with set label.
Guolin Ke's avatar
Guolin Ke committed
2651
2652
        """
        self.label = label
2653
        if self._handle is not None:
2654
2655
2656
2657
            if isinstance(label, pd_DataFrame):
                if len(label.columns) > 1:
                    raise ValueError('DataFrame for label cannot have multiple columns')
                _check_for_bad_pandas_dtypes(label.dtypes)
2658
2659
2660
2661
2662
2663
2664
2665
2666
2667
2668
                try:
                    # most common case (no nullable dtypes)
                    label = label.to_numpy(dtype=np.float32, copy=False)
                except TypeError:
                    # 1.0 <= pd version < 1.1 and nullable dtypes, least common case
                    # raises error because array is casted to type(pd.NA) and there's no na_value argument
                    label = label.astype(np.float32, copy=False).values
                except ValueError:
                    # data has nullable dtypes, but we can specify na_value argument and copy will be made
                    label = label.to_numpy(dtype=np.float32, na_value=np.nan)
                label_array = np.ravel(label)
2669
            else:
2670
                label_array = _list_to_1d_numpy(label, dtype=np.float32, name='label')
2671
            self.set_field('label', label_array)
2672
            self.label = self.get_field('label')  # original values can be modified at cpp side
Nikita Titov's avatar
Nikita Titov committed
2673
        return self
Guolin Ke's avatar
Guolin Ke committed
2674

2675
2676
2677
2678
    def set_weight(
        self,
        weight: Optional[_LGBM_WeightType]
    ) -> "Dataset":
2679
        """Set weight of each instance.
Guolin Ke's avatar
Guolin Ke committed
2680
2681
2682

        Parameters
        ----------
2683
        weight : list, numpy 1-D array, pandas Series or None
2684
            Weight to be set for each data point. Weights should be non-negative.
Nikita Titov's avatar
Nikita Titov committed
2685
2686
2687
2688
2689

        Returns
        -------
        self : Dataset
            Dataset with set weight.
Guolin Ke's avatar
Guolin Ke committed
2690
        """
2691
2692
        if weight is not None and np.all(weight == 1):
            weight = None
Guolin Ke's avatar
Guolin Ke committed
2693
        self.weight = weight
2694
        if self._handle is not None and weight is not None:
2695
            weight = _list_to_1d_numpy(weight, dtype=np.float32, name='weight')
wxchan's avatar
wxchan committed
2696
            self.set_field('weight', weight)
2697
            self.weight = self.get_field('weight')  # original values can be modified at cpp side
Nikita Titov's avatar
Nikita Titov committed
2698
        return self
Guolin Ke's avatar
Guolin Ke committed
2699

2700
2701
2702
2703
    def set_init_score(
        self,
        init_score: Optional[_LGBM_InitScoreType]
    ) -> "Dataset":
2704
        """Set init score of Booster to start from.
Guolin Ke's avatar
Guolin Ke committed
2705
2706
2707

        Parameters
        ----------
2708
        init_score : list, list of lists (for multi-class task), numpy array, pandas Series, pandas DataFrame (for multi-class task), or None
2709
            Init score for Booster.
Nikita Titov's avatar
Nikita Titov committed
2710
2711
2712
2713
2714

        Returns
        -------
        self : Dataset
            Dataset with set init score.
Guolin Ke's avatar
Guolin Ke committed
2715
2716
        """
        self.init_score = init_score
2717
        if self._handle is not None and init_score is not None:
wxchan's avatar
wxchan committed
2718
            self.set_field('init_score', init_score)
2719
            self.init_score = self.get_field('init_score')  # original values can be modified at cpp side
Nikita Titov's avatar
Nikita Titov committed
2720
        return self
Guolin Ke's avatar
Guolin Ke committed
2721

2722
2723
2724
2725
    def set_group(
        self,
        group: Optional[_LGBM_GroupType]
    ) -> "Dataset":
2726
        """Set group size of Dataset (used for ranking).
Guolin Ke's avatar
Guolin Ke committed
2727
2728
2729

        Parameters
        ----------
2730
        group : list, numpy 1-D array, pandas Series or None
2731
2732
2733
            Group/query data.
            Only used in the learning-to-rank task.
            sum(group) = n_samples.
2734
2735
            For example, if you have a 100-document dataset with ``group = [10, 20, 40, 10, 10, 10]``, that means that you have 6 groups,
            where the first 10 records are in the first group, records 11-30 are in the second group, records 31-70 are in the third group, etc.
Nikita Titov's avatar
Nikita Titov committed
2736
2737
2738
2739
2740

        Returns
        -------
        self : Dataset
            Dataset with set group.
Guolin Ke's avatar
Guolin Ke committed
2741
2742
        """
        self.group = group
2743
        if self._handle is not None and group is not None:
2744
            group = _list_to_1d_numpy(group, dtype=np.int32, name='group')
wxchan's avatar
wxchan committed
2745
            self.set_field('group', group)
Nikita Titov's avatar
Nikita Titov committed
2746
        return self
Guolin Ke's avatar
Guolin Ke committed
2747

2748
2749
2750
2751
2752
2753
2754
2755
2756
2757
2758
2759
2760
2761
2762
2763
2764
2765
2766
2767
2768
2769
    def set_position(
        self,
        position: Optional[_LGBM_PositionType]
    ) -> "Dataset":
        """Set position of Dataset (used for ranking).

        Parameters
        ----------
        position : numpy 1-D array, pandas Series or None, optional (default=None)
            Position of items used in unbiased learning-to-rank task.

        Returns
        -------
        self : Dataset
            Dataset with set position.
        """
        self.position = position
        if self._handle is not None and position is not None:
            position = _list_to_1d_numpy(position, dtype=np.int32, name='position')
            self.set_field('position', position)
        return self

2770
    def get_feature_name(self) -> List[str]:
2771
2772
2773
2774
        """Get the names of columns (features) in the Dataset.

        Returns
        -------
2775
        feature_names : list of str
2776
2777
            The names of columns (features) in the Dataset.
        """
2778
        if self._handle is None:
2779
2780
2781
2782
2783
            raise LightGBMError("Cannot get feature_name before construct dataset")
        num_feature = self.num_feature()
        tmp_out_len = ctypes.c_int(0)
        reserved_string_buffer_size = 255
        required_string_buffer_size = ctypes.c_size_t(0)
2784
        string_buffers = [ctypes.create_string_buffer(reserved_string_buffer_size) for _ in range(num_feature)]
2785
2786
        ptr_string_buffers = (ctypes.c_char_p * num_feature)(*map(ctypes.addressof, string_buffers))
        _safe_call(_LIB.LGBM_DatasetGetFeatureNames(
2787
            self._handle,
2788
            ctypes.c_int(num_feature),
2789
            ctypes.byref(tmp_out_len),
2790
            ctypes.c_size_t(reserved_string_buffer_size),
2791
2792
2793
2794
            ctypes.byref(required_string_buffer_size),
            ptr_string_buffers))
        if num_feature != tmp_out_len.value:
            raise ValueError("Length of feature names doesn't equal with num_feature")
2795
2796
2797
2798
2799
2800
        actual_string_buffer_size = required_string_buffer_size.value
        # if buffer length is not long enough, reallocate buffers
        if reserved_string_buffer_size < actual_string_buffer_size:
            string_buffers = [ctypes.create_string_buffer(actual_string_buffer_size) for _ in range(num_feature)]
            ptr_string_buffers = (ctypes.c_char_p * num_feature)(*map(ctypes.addressof, string_buffers))
            _safe_call(_LIB.LGBM_DatasetGetFeatureNames(
2801
                self._handle,
2802
2803
2804
2805
2806
                ctypes.c_int(num_feature),
                ctypes.byref(tmp_out_len),
                ctypes.c_size_t(actual_string_buffer_size),
                ctypes.byref(required_string_buffer_size),
                ptr_string_buffers))
2807
        return [string_buffers[i].value.decode('utf-8') for i in range(num_feature)]
2808

2809
    def get_label(self) -> Optional[np.ndarray]:
2810
        """Get the label of the Dataset.
Guolin Ke's avatar
Guolin Ke committed
2811
2812
2813

        Returns
        -------
Nikita Titov's avatar
Nikita Titov committed
2814
        label : numpy array or None
2815
            The label information from the Dataset.
Guolin Ke's avatar
Guolin Ke committed
2816
        """
2817
        if self.label is None:
wxchan's avatar
wxchan committed
2818
            self.label = self.get_field('label')
Guolin Ke's avatar
Guolin Ke committed
2819
2820
        return self.label

2821
    def get_weight(self) -> Optional[np.ndarray]:
2822
        """Get the weight of the Dataset.
Guolin Ke's avatar
Guolin Ke committed
2823
2824
2825

        Returns
        -------
Nikita Titov's avatar
Nikita Titov committed
2826
        weight : numpy array or None
2827
            Weight for each data point from the Dataset. Weights should be non-negative.
Guolin Ke's avatar
Guolin Ke committed
2828
        """
2829
        if self.weight is None:
wxchan's avatar
wxchan committed
2830
            self.weight = self.get_field('weight')
Guolin Ke's avatar
Guolin Ke committed
2831
2832
        return self.weight

2833
    def get_init_score(self) -> Optional[np.ndarray]:
2834
        """Get the initial score of the Dataset.
Guolin Ke's avatar
Guolin Ke committed
2835
2836
2837

        Returns
        -------
Nikita Titov's avatar
Nikita Titov committed
2838
        init_score : numpy array or None
2839
            Init score of Booster.
Guolin Ke's avatar
Guolin Ke committed
2840
        """
2841
        if self.init_score is None:
wxchan's avatar
wxchan committed
2842
            self.init_score = self.get_field('init_score')
Guolin Ke's avatar
Guolin Ke committed
2843
2844
        return self.init_score

2845
    def get_data(self) -> Optional[_LGBM_TrainDataType]:
2846
2847
2848
2849
        """Get the raw data of the Dataset.

        Returns
        -------
2850
        data : str, pathlib.Path, numpy array, pandas DataFrame, H2O DataTable's Frame, scipy.sparse, Sequence, list of Sequence or list of numpy array or None
2851
2852
            Raw data used in the Dataset construction.
        """
2853
        if self._handle is None:
2854
            raise Exception("Cannot get data before construct Dataset")
2855
        if self._need_slice and self.used_indices is not None and self.reference is not None:
Guolin Ke's avatar
Guolin Ke committed
2856
2857
            self.data = self.reference.data
            if self.data is not None:
2858
                if isinstance(self.data, np.ndarray) or isinstance(self.data, scipy.sparse.spmatrix):
Guolin Ke's avatar
Guolin Ke committed
2859
                    self.data = self.data[self.used_indices, :]
2860
                elif isinstance(self.data, pd_DataFrame):
Guolin Ke's avatar
Guolin Ke committed
2861
                    self.data = self.data.iloc[self.used_indices].copy()
2862
                elif isinstance(self.data, dt_DataTable):
Guolin Ke's avatar
Guolin Ke committed
2863
                    self.data = self.data[self.used_indices, :]
2864
2865
2866
                elif isinstance(self.data, Sequence):
                    self.data = self.data[self.used_indices]
                elif isinstance(self.data, list) and len(self.data) > 0 and all(isinstance(x, Sequence) for x in self.data):
2867
                    self.data = np.array(list(self._yield_row_from_seqlist(self.data, self.used_indices)))
Guolin Ke's avatar
Guolin Ke committed
2868
                else:
2869
2870
                    _log_warning(f"Cannot subset {type(self.data).__name__} type of raw data.\n"
                                 "Returning original raw data")
2871
            self._need_slice = False
Guolin Ke's avatar
Guolin Ke committed
2872
2873
2874
        if self.data is None:
            raise LightGBMError("Cannot call `get_data` after freed raw data, "
                                "set free_raw_data=False when construct Dataset to avoid this.")
2875
2876
        return self.data

2877
    def get_group(self) -> Optional[np.ndarray]:
2878
        """Get the group of the Dataset.
Guolin Ke's avatar
Guolin Ke committed
2879
2880
2881

        Returns
        -------
Nikita Titov's avatar
Nikita Titov committed
2882
        group : numpy array or None
2883
2884
2885
            Group/query data.
            Only used in the learning-to-rank task.
            sum(group) = n_samples.
2886
2887
            For example, if you have a 100-document dataset with ``group = [10, 20, 40, 10, 10, 10]``, that means that you have 6 groups,
            where the first 10 records are in the first group, records 11-30 are in the second group, records 31-70 are in the third group, etc.
Guolin Ke's avatar
Guolin Ke committed
2888
        """
2889
        if self.group is None:
wxchan's avatar
wxchan committed
2890
            self.group = self.get_field('group')
Guolin Ke's avatar
Guolin Ke committed
2891
2892
            if self.group is not None:
                # group data from LightGBM is boundaries data, need to convert to group size
Nikita Titov's avatar
Nikita Titov committed
2893
                self.group = np.diff(self.group)
Guolin Ke's avatar
Guolin Ke committed
2894
2895
        return self.group

2896
2897
2898
2899
2900
2901
2902
2903
2904
2905
2906
2907
    def get_position(self) -> Optional[np.ndarray]:
        """Get the position of the Dataset.

        Returns
        -------
        position : numpy 1-D array or None
            Position of items used in unbiased learning-to-rank task.
        """
        if self.position is None:
            self.position = self.get_field('position')
        return self.position

2908
    def num_data(self) -> int:
2909
        """Get the number of rows in the Dataset.
Guolin Ke's avatar
Guolin Ke committed
2910
2911
2912

        Returns
        -------
2913
2914
        number_of_rows : int
            The number of rows in the Dataset.
Guolin Ke's avatar
Guolin Ke committed
2915
        """
2916
        if self._handle is not None:
2917
            ret = ctypes.c_int(0)
2918
            _safe_call(_LIB.LGBM_DatasetGetNumData(self._handle,
wxchan's avatar
wxchan committed
2919
2920
                                                   ctypes.byref(ret)))
            return ret.value
Guolin Ke's avatar
Guolin Ke committed
2921
        else:
2922
            raise LightGBMError("Cannot get num_data before construct dataset")
Guolin Ke's avatar
Guolin Ke committed
2923

2924
    def num_feature(self) -> int:
2925
        """Get the number of columns (features) in the Dataset.
Guolin Ke's avatar
Guolin Ke committed
2926
2927
2928

        Returns
        -------
2929
2930
        number_of_columns : int
            The number of columns (features) in the Dataset.
Guolin Ke's avatar
Guolin Ke committed
2931
        """
2932
        if self._handle is not None:
2933
            ret = ctypes.c_int(0)
2934
            _safe_call(_LIB.LGBM_DatasetGetNumFeature(self._handle,
wxchan's avatar
wxchan committed
2935
2936
                                                      ctypes.byref(ret)))
            return ret.value
Guolin Ke's avatar
Guolin Ke committed
2937
        else:
2938
            raise LightGBMError("Cannot get num_feature before construct dataset")
Guolin Ke's avatar
Guolin Ke committed
2939

2940
    def feature_num_bin(self, feature: Union[int, str]) -> int:
2941
2942
        """Get the number of bins for a feature.

2943
2944
        .. versionadded:: 4.0.0

2945
2946
        Parameters
        ----------
2947
2948
        feature : int or str
            Index or name of the feature.
2949
2950
2951
2952
2953
2954

        Returns
        -------
        number_of_bins : int
            The number of constructed bins for the feature in the Dataset.
        """
2955
        if self._handle is not None:
2956
            if isinstance(feature, str):
2957
2958
2959
                feature_index = self.feature_name.index(feature)
            else:
                feature_index = feature
2960
            ret = ctypes.c_int(0)
2961
            _safe_call(_LIB.LGBM_DatasetGetFeatureNumBin(self._handle,
2962
                                                         ctypes.c_int(feature_index),
2963
2964
2965
2966
2967
                                                         ctypes.byref(ret)))
            return ret.value
        else:
            raise LightGBMError("Cannot get feature_num_bin before construct dataset")

2968
    def get_ref_chain(self, ref_limit: int = 100) -> Set["Dataset"]:
2969
2970
2971
2972
2973
        """Get a chain of Dataset objects.

        Starts with r, then goes to r.reference (if exists),
        then to r.reference.reference, etc.
        until we hit ``ref_limit`` or a reference loop.
2974
2975
2976
2977
2978

        Parameters
        ----------
        ref_limit : int, optional (default=100)
            The limit number of references.
2979
2980
2981

        Returns
        -------
2982
2983
2984
        ref_chain : set of Dataset
            Chain of references of the Datasets.
        """
2985
        head = self
2986
        ref_chain: Set[Dataset] = set()
2987
2988
        while len(ref_chain) < ref_limit:
            if isinstance(head, Dataset):
2989
                ref_chain.add(head)
2990
2991
2992
2993
2994
2995
                if (head.reference is not None) and (head.reference not in ref_chain):
                    head = head.reference
                else:
                    break
            else:
                break
Nikita Titov's avatar
Nikita Titov committed
2996
        return ref_chain
2997

2998
    def add_features_from(self, other: "Dataset") -> "Dataset":
2999
3000
3001
3002
3003
3004
3005
3006
3007
3008
3009
3010
3011
3012
        """Add features from other Dataset to the current Dataset.

        Both Datasets must be constructed before calling this method.

        Parameters
        ----------
        other : Dataset
            The Dataset to take features from.

        Returns
        -------
        self : Dataset
            Dataset with the new features added.
        """
3013
        if self._handle is None or other._handle is None:
3014
            raise ValueError('Both source and target Datasets must be constructed before adding features')
3015
        _safe_call(_LIB.LGBM_DatasetAddFeaturesFrom(self._handle, other._handle))
Guolin Ke's avatar
Guolin Ke committed
3016
3017
3018
3019
3020
3021
3022
3023
        was_none = self.data is None
        old_self_data_type = type(self.data).__name__
        if other.data is None:
            self.data = None
        elif self.data is not None:
            if isinstance(self.data, np.ndarray):
                if isinstance(other.data, np.ndarray):
                    self.data = np.hstack((self.data, other.data))
3024
                elif isinstance(other.data, scipy.sparse.spmatrix):
Guolin Ke's avatar
Guolin Ke committed
3025
                    self.data = np.hstack((self.data, other.data.toarray()))
3026
                elif isinstance(other.data, pd_DataFrame):
Guolin Ke's avatar
Guolin Ke committed
3027
                    self.data = np.hstack((self.data, other.data.values))
3028
                elif isinstance(other.data, dt_DataTable):
Guolin Ke's avatar
Guolin Ke committed
3029
3030
3031
                    self.data = np.hstack((self.data, other.data.to_numpy()))
                else:
                    self.data = None
3032
            elif isinstance(self.data, scipy.sparse.spmatrix):
Guolin Ke's avatar
Guolin Ke committed
3033
                sparse_format = self.data.getformat()
3034
                if isinstance(other.data, np.ndarray) or isinstance(other.data, scipy.sparse.spmatrix):
Guolin Ke's avatar
Guolin Ke committed
3035
                    self.data = scipy.sparse.hstack((self.data, other.data), format=sparse_format)
3036
                elif isinstance(other.data, pd_DataFrame):
Guolin Ke's avatar
Guolin Ke committed
3037
                    self.data = scipy.sparse.hstack((self.data, other.data.values), format=sparse_format)
3038
                elif isinstance(other.data, dt_DataTable):
Guolin Ke's avatar
Guolin Ke committed
3039
3040
3041
                    self.data = scipy.sparse.hstack((self.data, other.data.to_numpy()), format=sparse_format)
                else:
                    self.data = None
3042
            elif isinstance(self.data, pd_DataFrame):
Guolin Ke's avatar
Guolin Ke committed
3043
3044
                if not PANDAS_INSTALLED:
                    raise LightGBMError("Cannot add features to DataFrame type of raw data "
3045
3046
                                        "without pandas installed. "
                                        "Install pandas and restart your session.")
Guolin Ke's avatar
Guolin Ke committed
3047
                if isinstance(other.data, np.ndarray):
3048
                    self.data = concat((self.data, pd_DataFrame(other.data)),
Guolin Ke's avatar
Guolin Ke committed
3049
                                       axis=1, ignore_index=True)
3050
                elif isinstance(other.data, scipy.sparse.spmatrix):
3051
                    self.data = concat((self.data, pd_DataFrame(other.data.toarray())),
Guolin Ke's avatar
Guolin Ke committed
3052
                                       axis=1, ignore_index=True)
3053
                elif isinstance(other.data, pd_DataFrame):
Guolin Ke's avatar
Guolin Ke committed
3054
3055
                    self.data = concat((self.data, other.data),
                                       axis=1, ignore_index=True)
3056
3057
                elif isinstance(other.data, dt_DataTable):
                    self.data = concat((self.data, pd_DataFrame(other.data.to_numpy())),
Guolin Ke's avatar
Guolin Ke committed
3058
3059
3060
                                       axis=1, ignore_index=True)
                else:
                    self.data = None
3061
            elif isinstance(self.data, dt_DataTable):
Guolin Ke's avatar
Guolin Ke committed
3062
                if isinstance(other.data, np.ndarray):
3063
                    self.data = dt_DataTable(np.hstack((self.data.to_numpy(), other.data)))
3064
                elif isinstance(other.data, scipy.sparse.spmatrix):
3065
3066
3067
3068
3069
                    self.data = dt_DataTable(np.hstack((self.data.to_numpy(), other.data.toarray())))
                elif isinstance(other.data, pd_DataFrame):
                    self.data = dt_DataTable(np.hstack((self.data.to_numpy(), other.data.values)))
                elif isinstance(other.data, dt_DataTable):
                    self.data = dt_DataTable(np.hstack((self.data.to_numpy(), other.data.to_numpy())))
Guolin Ke's avatar
Guolin Ke committed
3070
3071
3072
3073
3074
                else:
                    self.data = None
            else:
                self.data = None
        if self.data is None:
3075
3076
            err_msg = (f"Cannot add features from {type(other.data).__name__} type of raw data to "
                       f"{old_self_data_type} type of raw data.\n")
Guolin Ke's avatar
Guolin Ke committed
3077
3078
            err_msg += ("Set free_raw_data=False when construct Dataset to avoid this"
                        if was_none else "Freeing raw data")
3079
            _log_warning(err_msg)
Guolin Ke's avatar
Guolin Ke committed
3080
        self.feature_name = self.get_feature_name()
3081
3082
        _log_warning("Reseting categorical features.\n"
                     "You can set new categorical features via ``set_categorical_feature`` method")
Guolin Ke's avatar
Guolin Ke committed
3083
3084
        self.categorical_feature = "auto"
        self.pandas_categorical = None
3085
3086
        return self

3087
    def _dump_text(self, filename: Union[str, Path]) -> "Dataset":
3088
3089
3090
3091
3092
3093
        """Save Dataset to a text file.

        This format cannot be loaded back in by LightGBM, but is useful for debugging purposes.

        Parameters
        ----------
3094
        filename : str or pathlib.Path
3095
3096
3097
3098
3099
3100
3101
3102
            Name of the output file.

        Returns
        -------
        self : Dataset
            Returns self.
        """
        _safe_call(_LIB.LGBM_DatasetDumpText(
3103
            self.construct()._handle,
3104
            _c_str(str(filename))))
3105
3106
        return self

wxchan's avatar
wxchan committed
3107

3108
3109
3110
3111
_LGBM_CustomObjectiveFunction = Callable[
    [np.ndarray, Dataset],
    Tuple[np.ndarray, np.ndarray]
]
3112
3113
3114
3115
3116
3117
3118
3119
3120
3121
_LGBM_CustomEvalFunction = Union[
    Callable[
        [np.ndarray, Dataset],
        _LGBM_EvalFunctionResultType
    ],
    Callable[
        [np.ndarray, Dataset],
        List[_LGBM_EvalFunctionResultType]
    ]
]
3122
3123


3124
class Booster:
3125
    """Booster in LightGBM."""
3126

3127
3128
3129
3130
3131
3132
3133
    def __init__(
        self,
        params: Optional[Dict[str, Any]] = None,
        train_set: Optional[Dataset] = None,
        model_file: Optional[Union[str, Path]] = None,
        model_str: Optional[str] = None
    ):
3134
        """Initialize the Booster.
wxchan's avatar
wxchan committed
3135
3136
3137

        Parameters
        ----------
Nikita Titov's avatar
Nikita Titov committed
3138
        params : dict or None, optional (default=None)
3139
3140
3141
            Parameters for Booster.
        train_set : Dataset or None, optional (default=None)
            Training dataset.
3142
        model_file : str, pathlib.Path or None, optional (default=None)
wxchan's avatar
wxchan committed
3143
            Path to the model file.
3144
        model_str : str or None, optional (default=None)
3145
            Model will be loaded from this string.
wxchan's avatar
wxchan committed
3146
        """
3147
        self._handle = ctypes.c_void_p()
3148
        self._network = False
wxchan's avatar
wxchan committed
3149
        self.__need_reload_eval_info = True
3150
        self._train_data_name = "training"
3151
        self.__set_objective_to_none = False
wxchan's avatar
wxchan committed
3152
        self.best_iteration = -1
3153
        self.best_score: _LGBM_BoosterBestScoreType = {}
3154
        params = {} if params is None else deepcopy(params)
wxchan's avatar
wxchan committed
3155
        if train_set is not None:
3156
            # Training task
wxchan's avatar
wxchan committed
3157
            if not isinstance(train_set, Dataset):
3158
                raise TypeError(f'Training data should be Dataset instance, met {type(train_set).__name__}')
3159
3160
3161
3162
3163
3164
3165
3166
3167
3168
3169
3170
3171
3172
3173
3174
3175
3176
3177
3178
3179
3180
3181
3182
3183
3184
3185
3186
3187
3188
3189
3190
3191
3192
            params = _choose_param_value(
                main_param_name="machines",
                params=params,
                default_value=None
            )
            # if "machines" is given, assume user wants to do distributed learning, and set up network
            if params["machines"] is None:
                params.pop("machines", None)
            else:
                machines = params["machines"]
                if isinstance(machines, str):
                    num_machines_from_machine_list = len(machines.split(','))
                elif isinstance(machines, (list, set)):
                    num_machines_from_machine_list = len(machines)
                    machines = ','.join(machines)
                else:
                    raise ValueError("Invalid machines in params.")

                params = _choose_param_value(
                    main_param_name="num_machines",
                    params=params,
                    default_value=num_machines_from_machine_list
                )
                params = _choose_param_value(
                    main_param_name="local_listen_port",
                    params=params,
                    default_value=12400
                )
                self.set_network(
                    machines=machines,
                    local_listen_port=params["local_listen_port"],
                    listen_time_out=params.get("time_out", 120),
                    num_machines=params["num_machines"]
                )
3193
            # construct booster object
3194
3195
3196
            train_set.construct()
            # copy the parameters from train_set
            params.update(train_set.get_params())
3197
            params_str = _param_dict_to_str(params)
wxchan's avatar
wxchan committed
3198
            _safe_call(_LIB.LGBM_BoosterCreate(
3199
                train_set._handle,
3200
                _c_str(params_str),
3201
                ctypes.byref(self._handle)))
3202
            # save reference to data
wxchan's avatar
wxchan committed
3203
            self.train_set = train_set
3204
3205
            self.valid_sets: List[Dataset] = []
            self.name_valid_sets: List[str] = []
wxchan's avatar
wxchan committed
3206
            self.__num_dataset = 1
Guolin Ke's avatar
Guolin Ke committed
3207
3208
            self.__init_predictor = train_set._predictor
            if self.__init_predictor is not None:
wxchan's avatar
wxchan committed
3209
                _safe_call(_LIB.LGBM_BoosterMerge(
3210
3211
                    self._handle,
                    self.__init_predictor._handle))
Guolin Ke's avatar
Guolin Ke committed
3212
            out_num_class = ctypes.c_int(0)
wxchan's avatar
wxchan committed
3213
            _safe_call(_LIB.LGBM_BoosterGetNumClasses(
3214
                self._handle,
wxchan's avatar
wxchan committed
3215
3216
                ctypes.byref(out_num_class)))
            self.__num_class = out_num_class.value
3217
            # buffer for inner predict
3218
            self.__inner_predict_buffer: List[Optional[np.ndarray]] = [None]
wxchan's avatar
wxchan committed
3219
3220
            self.__is_predicted_cur_iter = [False]
            self.__get_eval_info()
3221
            self.pandas_categorical = train_set.pandas_categorical
3222
            self.train_set_version = train_set.version
wxchan's avatar
wxchan committed
3223
        elif model_file is not None:
3224
            # Prediction task
Guolin Ke's avatar
Guolin Ke committed
3225
            out_num_iterations = ctypes.c_int(0)
wxchan's avatar
wxchan committed
3226
            _safe_call(_LIB.LGBM_BoosterCreateFromModelfile(
3227
                _c_str(str(model_file)),
wxchan's avatar
wxchan committed
3228
                ctypes.byref(out_num_iterations),
3229
                ctypes.byref(self._handle)))
Guolin Ke's avatar
Guolin Ke committed
3230
            out_num_class = ctypes.c_int(0)
wxchan's avatar
wxchan committed
3231
            _safe_call(_LIB.LGBM_BoosterGetNumClasses(
3232
                self._handle,
wxchan's avatar
wxchan committed
3233
3234
                ctypes.byref(out_num_class)))
            self.__num_class = out_num_class.value
3235
            self.pandas_categorical = _load_pandas_categorical(file_name=model_file)
3236
3237
3238
            if params:
                _log_warning('Ignoring params argument, using parameters from model file.')
            params = self._get_loaded_param()
3239
        elif model_str is not None:
3240
            self.model_from_string(model_str)
wxchan's avatar
wxchan committed
3241
        else:
3242
3243
            raise TypeError('Need at least one training dataset or model file or model string '
                            'to create Booster instance')
3244
        self.params = params
wxchan's avatar
wxchan committed
3245

3246
    def __del__(self) -> None:
3247
        try:
3248
            if self._network:
3249
3250
3251
3252
                self.free_network()
        except AttributeError:
            pass
        try:
3253
3254
            if self._handle is not None:
                _safe_call(_LIB.LGBM_BoosterFree(self._handle))
3255
3256
        except AttributeError:
            pass
wxchan's avatar
wxchan committed
3257

3258
    def __copy__(self) -> "Booster":
wxchan's avatar
wxchan committed
3259
3260
        return self.__deepcopy__(None)

3261
    def __deepcopy__(self, _) -> "Booster":
3262
        model_str = self.model_to_string(num_iteration=-1)
3263
        return Booster(model_str=model_str)
wxchan's avatar
wxchan committed
3264

3265
    def __getstate__(self) -> Dict[str, Any]:
wxchan's avatar
wxchan committed
3266
        this = self.__dict__.copy()
3267
        handle = this['_handle']
wxchan's avatar
wxchan committed
3268
3269
3270
        this.pop('train_set', None)
        this.pop('valid_sets', None)
        if handle is not None:
3271
            this["_handle"] = self.model_to_string(num_iteration=-1)
wxchan's avatar
wxchan committed
3272
3273
        return this

3274
    def __setstate__(self, state: Dict[str, Any]) -> None:
3275
        model_str = state.get('_handle', state.get('handle', None))
3276
        if model_str is not None:
wxchan's avatar
wxchan committed
3277
            handle = ctypes.c_void_p()
Guolin Ke's avatar
Guolin Ke committed
3278
            out_num_iterations = ctypes.c_int(0)
3279
            _safe_call(_LIB.LGBM_BoosterLoadModelFromString(
3280
                _c_str(model_str),
3281
3282
                ctypes.byref(out_num_iterations),
                ctypes.byref(handle)))
3283
            state['_handle'] = handle
wxchan's avatar
wxchan committed
3284
3285
        self.__dict__.update(state)

3286
3287
3288
3289
3290
3291
    def _get_loaded_param(self) -> Dict[str, Any]:
        buffer_len = 1 << 20
        tmp_out_len = ctypes.c_int64(0)
        string_buffer = ctypes.create_string_buffer(buffer_len)
        ptr_string_buffer = ctypes.c_char_p(*[ctypes.addressof(string_buffer)])
        _safe_call(_LIB.LGBM_BoosterGetLoadedParam(
3292
            self._handle,
3293
3294
3295
3296
3297
3298
3299
3300
3301
            ctypes.c_int64(buffer_len),
            ctypes.byref(tmp_out_len),
            ptr_string_buffer))
        actual_len = tmp_out_len.value
        # if buffer length is not long enough, re-allocate a buffer
        if actual_len > buffer_len:
            string_buffer = ctypes.create_string_buffer(actual_len)
            ptr_string_buffer = ctypes.c_char_p(*[ctypes.addressof(string_buffer)])
            _safe_call(_LIB.LGBM_BoosterGetLoadedParam(
3302
                self._handle,
3303
3304
3305
3306
3307
                ctypes.c_int64(actual_len),
                ctypes.byref(tmp_out_len),
                ptr_string_buffer))
        return json.loads(string_buffer.value.decode('utf-8'))

3308
    def free_dataset(self) -> "Booster":
Nikita Titov's avatar
Nikita Titov committed
3309
3310
3311
3312
3313
3314
3315
        """Free Booster's Datasets.

        Returns
        -------
        self : Booster
            Booster without Datasets.
        """
wxchan's avatar
wxchan committed
3316
3317
        self.__dict__.pop('train_set', None)
        self.__dict__.pop('valid_sets', None)
3318
        self.__num_dataset = 0
Nikita Titov's avatar
Nikita Titov committed
3319
        return self
wxchan's avatar
wxchan committed
3320

3321
    def _free_buffer(self) -> "Booster":
3322
3323
        self.__inner_predict_buffer = []
        self.__is_predicted_cur_iter = []
Nikita Titov's avatar
Nikita Titov committed
3324
        return self
3325

3326
3327
3328
3329
3330
3331
3332
    def set_network(
        self,
        machines: Union[List[str], Set[str], str],
        local_listen_port: int = 12400,
        listen_time_out: int = 120,
        num_machines: int = 1
    ) -> "Booster":
3333
3334
3335
3336
        """Set the network configuration.

        Parameters
        ----------
3337
        machines : list, set or str
3338
            Names of machines.
Nikita Titov's avatar
Nikita Titov committed
3339
        local_listen_port : int, optional (default=12400)
3340
            TCP listen port for local machines.
Nikita Titov's avatar
Nikita Titov committed
3341
        listen_time_out : int, optional (default=120)
3342
            Socket time-out in minutes.
Nikita Titov's avatar
Nikita Titov committed
3343
        num_machines : int, optional (default=1)
3344
            The number of machines for distributed learning application.
Nikita Titov's avatar
Nikita Titov committed
3345
3346
3347
3348
3349

        Returns
        -------
        self : Booster
            Booster with set network.
3350
        """
3351
3352
        if isinstance(machines, (list, set)):
            machines = ','.join(machines)
3353
        _safe_call(_LIB.LGBM_NetworkInit(_c_str(machines),
3354
3355
3356
                                         ctypes.c_int(local_listen_port),
                                         ctypes.c_int(listen_time_out),
                                         ctypes.c_int(num_machines)))
3357
        self._network = True
Nikita Titov's avatar
Nikita Titov committed
3358
        return self
3359

3360
    def free_network(self) -> "Booster":
Nikita Titov's avatar
Nikita Titov committed
3361
3362
3363
3364
3365
3366
3367
        """Free Booster's network.

        Returns
        -------
        self : Booster
            Booster with freed network.
        """
3368
        _safe_call(_LIB.LGBM_NetworkFree())
3369
        self._network = False
Nikita Titov's avatar
Nikita Titov committed
3370
        return self
3371

3372
    def trees_to_dataframe(self) -> pd_DataFrame:
3373
3374
        """Parse the fitted model and return in an easy-to-read pandas DataFrame.

3375
3376
3377
3378
        The returned DataFrame has the following columns.

            - ``tree_index`` : int64, which tree a node belongs to. 0-based, so a value of ``6``, for example, means "this node is in the 7th tree".
            - ``node_depth`` : int64, how far a node is from the root of the tree. The root node has a value of ``1``, its direct children are ``2``, etc.
3379
3380
3381
3382
3383
            - ``node_index`` : str, unique identifier for a node.
            - ``left_child`` : str, ``node_index`` of the child node to the left of a split. ``None`` for leaf nodes.
            - ``right_child`` : str, ``node_index`` of the child node to the right of a split. ``None`` for leaf nodes.
            - ``parent_index`` : str, ``node_index`` of this node's parent. ``None`` for the root node.
            - ``split_feature`` : str, name of the feature used for splitting. ``None`` for leaf nodes.
3384
3385
            - ``split_gain`` : float64, gain from adding this split to the tree. ``NaN`` for leaf nodes.
            - ``threshold`` : float64, value of the feature used to decide which side of the split a record will go down. ``NaN`` for leaf nodes.
3386
            - ``decision_type`` : str, logical operator describing how to compare a value to ``threshold``.
3387
3388
              For example, ``split_feature = "Column_10", threshold = 15, decision_type = "<="`` means that
              records where ``Column_10 <= 15`` follow the left side of the split, otherwise follows the right side of the split. ``None`` for leaf nodes.
3389
3390
            - ``missing_direction`` : str, split direction that missing values should go to. ``None`` for leaf nodes.
            - ``missing_type`` : str, describes what types of values are treated as missing.
3391
            - ``value`` : float64, predicted value for this leaf node, multiplied by the learning rate.
3392
            - ``weight`` : float64 or int64, sum of Hessian (second-order derivative of objective), summed over observations that fall in this node.
3393
3394
            - ``count`` : int64, number of records in the training data that fall into this node.

3395
3396
3397
3398
3399
3400
        Returns
        -------
        result : pandas DataFrame
            Returns a pandas DataFrame of the parsed model.
        """
        if not PANDAS_INSTALLED:
3401
3402
            raise LightGBMError('This method cannot be run without pandas installed. '
                                'You must install pandas and restart your session to use this method.')
3403
3404
3405
3406

        if self.num_trees() == 0:
            raise LightGBMError('There are no trees in this Booster and thus nothing to parse')

3407
        def _is_split_node(tree: Dict[str, Any]) -> bool:
3408
3409
            return 'split_index' in tree.keys()

3410
3411
3412
3413
3414
3415
3416
3417
3418
3419
3420
3421
        def create_node_record(
            tree: Dict[str, Any],
            node_depth: int = 1,
            tree_index: Optional[int] = None,
            feature_names: Optional[List[str]] = None,
            parent_node: Optional[str] = None
        ) -> Dict[str, Any]:

            def _get_node_index(
                tree: Dict[str, Any],
                tree_index: Optional[int]
            ) -> str:
3422
                tree_num = f'{tree_index}-' if tree_index is not None else ''
3423
3424
3425
                is_split = _is_split_node(tree)
                node_type = 'S' if is_split else 'L'
                # if a single node tree it won't have `leaf_index` so return 0
3426
3427
                node_num = tree.get('split_index' if is_split else 'leaf_index', 0)
                return f"{tree_num}{node_type}{node_num}"
3428

3429
3430
3431
3432
            def _get_split_feature(
                tree: Dict[str, Any],
                feature_names: Optional[List[str]]
            ) -> Optional[str]:
3433
3434
3435
3436
3437
3438
3439
3440
3441
                if _is_split_node(tree):
                    if feature_names is not None:
                        feature_name = feature_names[tree['split_feature']]
                    else:
                        feature_name = tree['split_feature']
                else:
                    feature_name = None
                return feature_name

3442
            def _is_single_node_tree(tree: Dict[str, Any]) -> bool:
3443
                return set(tree.keys()) == {'leaf_value'}
3444
3445

            # Create the node record, and populate universal data members
3446
            node: Dict[str, Union[int, str, None]] = OrderedDict()
3447
3448
3449
3450
3451
3452
3453
3454
3455
3456
3457
3458
3459
3460
3461
3462
3463
3464
3465
3466
3467
3468
3469
3470
3471
3472
3473
3474
3475
3476
3477
3478
3479
3480
3481
3482
            node['tree_index'] = tree_index
            node['node_depth'] = node_depth
            node['node_index'] = _get_node_index(tree, tree_index)
            node['left_child'] = None
            node['right_child'] = None
            node['parent_index'] = parent_node
            node['split_feature'] = _get_split_feature(tree, feature_names)
            node['split_gain'] = None
            node['threshold'] = None
            node['decision_type'] = None
            node['missing_direction'] = None
            node['missing_type'] = None
            node['value'] = None
            node['weight'] = None
            node['count'] = None

            # Update values to reflect node type (leaf or split)
            if _is_split_node(tree):
                node['left_child'] = _get_node_index(tree['left_child'], tree_index)
                node['right_child'] = _get_node_index(tree['right_child'], tree_index)
                node['split_gain'] = tree['split_gain']
                node['threshold'] = tree['threshold']
                node['decision_type'] = tree['decision_type']
                node['missing_direction'] = 'left' if tree['default_left'] else 'right'
                node['missing_type'] = tree['missing_type']
                node['value'] = tree['internal_value']
                node['weight'] = tree['internal_weight']
                node['count'] = tree['internal_count']
            else:
                node['value'] = tree['leaf_value']
                if not _is_single_node_tree(tree):
                    node['weight'] = tree['leaf_weight']
                    node['count'] = tree['leaf_count']

            return node

3483
3484
3485
3486
3487
3488
3489
        def tree_dict_to_node_list(
            tree: Dict[str, Any],
            node_depth: int = 1,
            tree_index: Optional[int] = None,
            feature_names: Optional[List[str]] = None,
            parent_node: Optional[str] = None
        ) -> List[Dict[str, Any]]:
3490

3491
            node = create_node_record(tree=tree,
3492
3493
3494
3495
3496
3497
3498
3499
3500
3501
3502
3503
                                      node_depth=node_depth,
                                      tree_index=tree_index,
                                      feature_names=feature_names,
                                      parent_node=parent_node)

            res = [node]

            if _is_split_node(tree):
                # traverse the next level of the tree
                children = ['left_child', 'right_child']
                for child in children:
                    subtree_list = tree_dict_to_node_list(
3504
                        tree=tree[child],
3505
3506
3507
                        node_depth=node_depth + 1,
                        tree_index=tree_index,
                        feature_names=feature_names,
3508
3509
                        parent_node=node['node_index']
                    )
3510
3511
3512
3513
3514
3515
3516
3517
3518
                    # In tree format, "subtree_list" is a list of node records (dicts),
                    # and we add node to the list.
                    res.extend(subtree_list)
            return res

        model_dict = self.dump_model()
        feature_names = model_dict['feature_names']
        model_list = []
        for tree in model_dict['tree_info']:
3519
            model_list.extend(tree_dict_to_node_list(tree=tree['tree_structure'],
3520
3521
3522
                                                     tree_index=tree['tree_index'],
                                                     feature_names=feature_names))

3523
        return pd_DataFrame(model_list, columns=model_list[0].keys())
3524

3525
    def set_train_data_name(self, name: str) -> "Booster":
3526
3527
3528
3529
        """Set the name to the training Dataset.

        Parameters
        ----------
3530
        name : str
Nikita Titov's avatar
Nikita Titov committed
3531
3532
3533
3534
3535
3536
            Name for the training Dataset.

        Returns
        -------
        self : Booster
            Booster with set training Dataset name.
3537
        """
3538
        self._train_data_name = name
Nikita Titov's avatar
Nikita Titov committed
3539
        return self
wxchan's avatar
wxchan committed
3540

3541
    def add_valid(self, data: Dataset, name: str) -> "Booster":
3542
        """Add validation data.
wxchan's avatar
wxchan committed
3543
3544
3545
3546

        Parameters
        ----------
        data : Dataset
3547
            Validation data.
3548
        name : str
3549
            Name of validation data.
Nikita Titov's avatar
Nikita Titov committed
3550
3551
3552
3553
3554

        Returns
        -------
        self : Booster
            Booster with set validation data.
wxchan's avatar
wxchan committed
3555
        """
Guolin Ke's avatar
Guolin Ke committed
3556
        if not isinstance(data, Dataset):
3557
            raise TypeError(f'Validation data should be Dataset instance, met {type(data).__name__}')
Guolin Ke's avatar
Guolin Ke committed
3558
        if data._predictor is not self.__init_predictor:
3559
3560
            raise LightGBMError("Add validation data failed, "
                                "you should use same predictor for these data")
wxchan's avatar
wxchan committed
3561
        _safe_call(_LIB.LGBM_BoosterAddValidData(
3562
3563
            self._handle,
            data.construct()._handle))
wxchan's avatar
wxchan committed
3564
3565
3566
3567
3568
        self.valid_sets.append(data)
        self.name_valid_sets.append(name)
        self.__num_dataset += 1
        self.__inner_predict_buffer.append(None)
        self.__is_predicted_cur_iter.append(False)
Nikita Titov's avatar
Nikita Titov committed
3569
        return self
wxchan's avatar
wxchan committed
3570

3571
    def reset_parameter(self, params: Dict[str, Any]) -> "Booster":
3572
        """Reset parameters of Booster.
wxchan's avatar
wxchan committed
3573
3574
3575
3576

        Parameters
        ----------
        params : dict
3577
            New parameters for Booster.
Nikita Titov's avatar
Nikita Titov committed
3578
3579
3580
3581
3582

        Returns
        -------
        self : Booster
            Booster with new parameters.
wxchan's avatar
wxchan committed
3583
        """
3584
        params_str = _param_dict_to_str(params)
wxchan's avatar
wxchan committed
3585
3586
        if params_str:
            _safe_call(_LIB.LGBM_BoosterResetParameter(
3587
                self._handle,
3588
                _c_str(params_str)))
Guolin Ke's avatar
Guolin Ke committed
3589
        self.params.update(params)
Nikita Titov's avatar
Nikita Titov committed
3590
        return self
wxchan's avatar
wxchan committed
3591

3592
3593
3594
3595
3596
    def update(
        self,
        train_set: Optional[Dataset] = None,
        fobj: Optional[_LGBM_CustomObjectiveFunction] = None
    ) -> bool:
Nikita Titov's avatar
Nikita Titov committed
3597
        """Update Booster for one iteration.
3598

wxchan's avatar
wxchan committed
3599
3600
        Parameters
        ----------
3601
3602
3603
3604
        train_set : Dataset or None, optional (default=None)
            Training data.
            If None, last training data is used.
        fobj : callable or None, optional (default=None)
wxchan's avatar
wxchan committed
3605
            Customized objective function.
3606
3607
3608
            Should accept two parameters: preds, train_data,
            and return (grad, hess).

3609
                preds : numpy 1-D array or numpy 2-D array (for multi-class task)
3610
                    The predicted values.
3611
3612
                    Predicted values are returned before any transformation,
                    e.g. they are raw margin instead of probability of positive class for binary task.
3613
3614
                train_data : Dataset
                    The training dataset.
3615
                grad : numpy 1-D array or numpy 2-D array (for multi-class task)
3616
3617
                    The value of the first order derivative (gradient) of the loss
                    with respect to the elements of preds for each sample point.
3618
                hess : numpy 1-D array or numpy 2-D array (for multi-class task)
3619
3620
                    The value of the second order derivative (Hessian) of the loss
                    with respect to the elements of preds for each sample point.
wxchan's avatar
wxchan committed
3621

3622
            For multi-class task, preds are numpy 2-D array of shape = [n_samples, n_classes],
3623
            and grad and hess should be returned in the same format.
3624

wxchan's avatar
wxchan committed
3625
3626
        Returns
        -------
3627
3628
        is_finished : bool
            Whether the update was successfully finished.
wxchan's avatar
wxchan committed
3629
        """
3630
        # need reset training data
3631
3632
3633
3634
3635
3636
        if train_set is None and self.train_set_version != self.train_set.version:
            train_set = self.train_set
            is_the_same_train_set = False
        else:
            is_the_same_train_set = train_set is self.train_set and self.train_set_version == train_set.version
        if train_set is not None and not is_the_same_train_set:
Guolin Ke's avatar
Guolin Ke committed
3637
            if not isinstance(train_set, Dataset):
3638
                raise TypeError(f'Training data should be Dataset instance, met {type(train_set).__name__}')
Guolin Ke's avatar
Guolin Ke committed
3639
            if train_set._predictor is not self.__init_predictor:
3640
3641
                raise LightGBMError("Replace training data failed, "
                                    "you should use same predictor for these data")
wxchan's avatar
wxchan committed
3642
3643
            self.train_set = train_set
            _safe_call(_LIB.LGBM_BoosterResetTrainingData(
3644
3645
                self._handle,
                self.train_set.construct()._handle))
wxchan's avatar
wxchan committed
3646
            self.__inner_predict_buffer[0] = None
3647
            self.train_set_version = self.train_set.version
wxchan's avatar
wxchan committed
3648
3649
        is_finished = ctypes.c_int(0)
        if fobj is None:
3650
            if self.__set_objective_to_none:
3651
                raise LightGBMError('Cannot update due to null objective function.')
wxchan's avatar
wxchan committed
3652
            _safe_call(_LIB.LGBM_BoosterUpdateOneIter(
3653
                self._handle,
wxchan's avatar
wxchan committed
3654
                ctypes.byref(is_finished)))
3655
            self.__is_predicted_cur_iter = [False for _ in range(self.__num_dataset)]
wxchan's avatar
wxchan committed
3656
3657
            return is_finished.value == 1
        else:
3658
            if not self.__set_objective_to_none:
Nikita Titov's avatar
Nikita Titov committed
3659
                self.reset_parameter({"objective": "none"}).__set_objective_to_none = True
wxchan's avatar
wxchan committed
3660
3661
3662
            grad, hess = fobj(self.__inner_predict(0), self.train_set)
            return self.__boost(grad, hess)

3663
3664
3665
3666
3667
    def __boost(
        self,
        grad: np.ndarray,
        hess: np.ndarray
    ) -> bool:
3668
        """Boost Booster for one iteration with customized gradient statistics.
Nikita Titov's avatar
Nikita Titov committed
3669

Nikita Titov's avatar
Nikita Titov committed
3670
3671
        .. note::

3672
3673
            Score is returned before any transformation,
            e.g. it is raw margin instead of probability of positive class for binary task.
3674
            For multi-class task, score are numpy 2-D array of shape = [n_samples, n_classes],
3675
            and grad and hess should be returned in the same format.
3676

wxchan's avatar
wxchan committed
3677
3678
        Parameters
        ----------
3679
        grad : numpy 1-D array or numpy 2-D array (for multi-class task)
3680
3681
            The value of the first order derivative (gradient) of the loss
            with respect to the elements of score for each sample point.
3682
        hess : numpy 1-D array or numpy 2-D array (for multi-class task)
3683
3684
            The value of the second order derivative (Hessian) of the loss
            with respect to the elements of score for each sample point.
wxchan's avatar
wxchan committed
3685
3686
3687

        Returns
        -------
Nikita Titov's avatar
Nikita Titov committed
3688
3689
        is_finished : bool
            Whether the boost was successfully finished.
wxchan's avatar
wxchan committed
3690
        """
3691
3692
3693
        if self.__num_class > 1:
            grad = grad.ravel(order='F')
            hess = hess.ravel(order='F')
3694
3695
        grad = _list_to_1d_numpy(grad, dtype=np.float32, name='gradient')
        hess = _list_to_1d_numpy(hess, dtype=np.float32, name='hessian')
3696
3697
        assert grad.flags.c_contiguous
        assert hess.flags.c_contiguous
wxchan's avatar
wxchan committed
3698
        if len(grad) != len(hess):
3699
3700
            raise ValueError(f"Lengths of gradient ({len(grad)}) and Hessian ({len(hess)}) don't match")
        num_train_data = self.train_set.num_data()
3701
        if len(grad) != num_train_data * self.__num_class:
3702
3703
3704
            raise ValueError(
                f"Lengths of gradient ({len(grad)}) and Hessian ({len(hess)}) "
                f"don't match training data length ({num_train_data}) * "
3705
                f"number of models per one iteration ({self.__num_class})"
3706
            )
wxchan's avatar
wxchan committed
3707
3708
        is_finished = ctypes.c_int(0)
        _safe_call(_LIB.LGBM_BoosterUpdateOneIterCustom(
3709
            self._handle,
wxchan's avatar
wxchan committed
3710
3711
3712
            grad.ctypes.data_as(ctypes.POINTER(ctypes.c_float)),
            hess.ctypes.data_as(ctypes.POINTER(ctypes.c_float)),
            ctypes.byref(is_finished)))
3713
        self.__is_predicted_cur_iter = [False for _ in range(self.__num_dataset)]
wxchan's avatar
wxchan committed
3714
3715
        return is_finished.value == 1

3716
    def rollback_one_iter(self) -> "Booster":
Nikita Titov's avatar
Nikita Titov committed
3717
3718
3719
3720
3721
3722
3723
        """Rollback one iteration.

        Returns
        -------
        self : Booster
            Booster with rolled back one iteration.
        """
wxchan's avatar
wxchan committed
3724
        _safe_call(_LIB.LGBM_BoosterRollbackOneIter(
3725
            self._handle))
3726
        self.__is_predicted_cur_iter = [False for _ in range(self.__num_dataset)]
Nikita Titov's avatar
Nikita Titov committed
3727
        return self
wxchan's avatar
wxchan committed
3728

3729
    def current_iteration(self) -> int:
3730
3731
3732
3733
3734
3735
3736
        """Get the index of the current iteration.

        Returns
        -------
        cur_iter : int
            The index of the current iteration.
        """
Guolin Ke's avatar
Guolin Ke committed
3737
        out_cur_iter = ctypes.c_int(0)
wxchan's avatar
wxchan committed
3738
        _safe_call(_LIB.LGBM_BoosterGetCurrentIteration(
3739
            self._handle,
wxchan's avatar
wxchan committed
3740
3741
3742
            ctypes.byref(out_cur_iter)))
        return out_cur_iter.value

3743
    def num_model_per_iteration(self) -> int:
3744
3745
3746
3747
3748
3749
3750
3751
3752
        """Get number of models per iteration.

        Returns
        -------
        model_per_iter : int
            The number of models per iteration.
        """
        model_per_iter = ctypes.c_int(0)
        _safe_call(_LIB.LGBM_BoosterNumModelPerIteration(
3753
            self._handle,
3754
3755
3756
            ctypes.byref(model_per_iter)))
        return model_per_iter.value

3757
    def num_trees(self) -> int:
3758
3759
3760
3761
3762
3763
3764
3765
3766
        """Get number of weak sub-models.

        Returns
        -------
        num_trees : int
            The number of weak sub-models.
        """
        num_trees = ctypes.c_int(0)
        _safe_call(_LIB.LGBM_BoosterNumberOfTotalModel(
3767
            self._handle,
3768
3769
3770
            ctypes.byref(num_trees)))
        return num_trees.value

3771
    def upper_bound(self) -> float:
3772
3773
3774
3775
        """Get upper bound value of a model.

        Returns
        -------
3776
        upper_bound : float
3777
3778
3779
3780
            Upper bound value of the model.
        """
        ret = ctypes.c_double(0)
        _safe_call(_LIB.LGBM_BoosterGetUpperBoundValue(
3781
            self._handle,
3782
3783
3784
            ctypes.byref(ret)))
        return ret.value

3785
    def lower_bound(self) -> float:
3786
3787
3788
3789
        """Get lower bound value of a model.

        Returns
        -------
3790
        lower_bound : float
3791
3792
3793
3794
            Lower bound value of the model.
        """
        ret = ctypes.c_double(0)
        _safe_call(_LIB.LGBM_BoosterGetLowerBoundValue(
3795
            self._handle,
3796
3797
3798
            ctypes.byref(ret)))
        return ret.value

3799
3800
3801
3802
3803
3804
    def eval(
        self,
        data: Dataset,
        name: str,
        feval: Optional[Union[_LGBM_CustomEvalFunction, List[_LGBM_CustomEvalFunction]]] = None
    ) -> List[_LGBM_BoosterEvalMethodResultType]:
3805
        """Evaluate for data.
wxchan's avatar
wxchan committed
3806
3807
3808

        Parameters
        ----------
3809
3810
        data : Dataset
            Data for the evaluating.
3811
        name : str
3812
            Name of the data.
3813
        feval : callable, list of callable, or None, optional (default=None)
3814
            Customized evaluation function.
3815
            Each evaluation function should accept two parameters: preds, eval_data,
3816
            and return (eval_name, eval_result, is_higher_better) or list of such tuples.
3817

3818
                preds : numpy 1-D array or numpy 2-D array (for multi-class task)
3819
                    The predicted values.
3820
                    For multi-class task, preds are numpy 2-D array of shape = [n_samples, n_classes].
3821
                    If custom objective function is used, predicted values are returned before any transformation,
3822
                    e.g. they are raw margin instead of probability of positive class for binary task in this case.
3823
                eval_data : Dataset
3824
                    A ``Dataset`` to evaluate.
3825
                eval_name : str
Andrew Ziem's avatar
Andrew Ziem committed
3826
                    The name of evaluation function (without whitespace).
3827
3828
3829
3830
3831
                eval_result : float
                    The eval result.
                is_higher_better : bool
                    Is eval result higher better, e.g. AUC is ``is_higher_better``.

wxchan's avatar
wxchan committed
3832
3833
        Returns
        -------
Nikita Titov's avatar
Nikita Titov committed
3834
        result : list
3835
            List with (dataset_name, eval_name, eval_result, is_higher_better) tuples.
wxchan's avatar
wxchan committed
3836
        """
Guolin Ke's avatar
Guolin Ke committed
3837
3838
        if not isinstance(data, Dataset):
            raise TypeError("Can only eval for Dataset instance")
wxchan's avatar
wxchan committed
3839
3840
3841
3842
        data_idx = -1
        if data is self.train_set:
            data_idx = 0
        else:
3843
            for i in range(len(self.valid_sets)):
wxchan's avatar
wxchan committed
3844
3845
3846
                if data is self.valid_sets[i]:
                    data_idx = i + 1
                    break
3847
        # need to push new valid data
wxchan's avatar
wxchan committed
3848
3849
3850
3851
3852
3853
        if data_idx == -1:
            self.add_valid(data, name)
            data_idx = self.__num_dataset - 1

        return self.__inner_eval(name, data_idx, feval)

3854
3855
3856
3857
    def eval_train(
        self,
        feval: Optional[Union[_LGBM_CustomEvalFunction, List[_LGBM_CustomEvalFunction]]] = None
    ) -> List[_LGBM_BoosterEvalMethodResultType]:
3858
        """Evaluate for training data.
wxchan's avatar
wxchan committed
3859
3860
3861

        Parameters
        ----------
3862
        feval : callable, list of callable, or None, optional (default=None)
3863
            Customized evaluation function.
3864
            Each evaluation function should accept two parameters: preds, eval_data,
3865
            and return (eval_name, eval_result, is_higher_better) or list of such tuples.
3866

3867
                preds : numpy 1-D array or numpy 2-D array (for multi-class task)
3868
                    The predicted values.
3869
                    For multi-class task, preds are numpy 2-D array of shape = [n_samples, n_classes].
3870
                    If custom objective function is used, predicted values are returned before any transformation,
3871
                    e.g. they are raw margin instead of probability of positive class for binary task in this case.
Akshita Dixit's avatar
Akshita Dixit committed
3872
                eval_data : Dataset
3873
                    The training dataset.
3874
                eval_name : str
Andrew Ziem's avatar
Andrew Ziem committed
3875
                    The name of evaluation function (without whitespace).
3876
3877
3878
3879
3880
                eval_result : float
                    The eval result.
                is_higher_better : bool
                    Is eval result higher better, e.g. AUC is ``is_higher_better``.

wxchan's avatar
wxchan committed
3881
3882
        Returns
        -------
Nikita Titov's avatar
Nikita Titov committed
3883
        result : list
3884
            List with (train_dataset_name, eval_name, eval_result, is_higher_better) tuples.
wxchan's avatar
wxchan committed
3885
        """
3886
        return self.__inner_eval(self._train_data_name, 0, feval)
wxchan's avatar
wxchan committed
3887

3888
3889
3890
3891
    def eval_valid(
        self,
        feval: Optional[Union[_LGBM_CustomEvalFunction, List[_LGBM_CustomEvalFunction]]] = None
    ) -> List[_LGBM_BoosterEvalMethodResultType]:
3892
        """Evaluate for validation data.
wxchan's avatar
wxchan committed
3893
3894
3895

        Parameters
        ----------
3896
        feval : callable, list of callable, or None, optional (default=None)
3897
            Customized evaluation function.
3898
            Each evaluation function should accept two parameters: preds, eval_data,
3899
            and return (eval_name, eval_result, is_higher_better) or list of such tuples.
3900

3901
                preds : numpy 1-D array or numpy 2-D array (for multi-class task)
3902
                    The predicted values.
3903
                    For multi-class task, preds are numpy 2-D array of shape = [n_samples, n_classes].
3904
                    If custom objective function is used, predicted values are returned before any transformation,
3905
                    e.g. they are raw margin instead of probability of positive class for binary task in this case.
Akshita Dixit's avatar
Akshita Dixit committed
3906
                eval_data : Dataset
3907
                    The validation dataset.
3908
                eval_name : str
Andrew Ziem's avatar
Andrew Ziem committed
3909
                    The name of evaluation function (without whitespace).
3910
3911
3912
3913
3914
                eval_result : float
                    The eval result.
                is_higher_better : bool
                    Is eval result higher better, e.g. AUC is ``is_higher_better``.

wxchan's avatar
wxchan committed
3915
3916
        Returns
        -------
Nikita Titov's avatar
Nikita Titov committed
3917
        result : list
3918
            List with (validation_dataset_name, eval_name, eval_result, is_higher_better) tuples.
wxchan's avatar
wxchan committed
3919
        """
3920
        return [item for i in range(1, self.__num_dataset)
wxchan's avatar
wxchan committed
3921
                for item in self.__inner_eval(self.name_valid_sets[i - 1], i, feval)]
wxchan's avatar
wxchan committed
3922

3923
3924
3925
3926
3927
3928
3929
    def save_model(
        self,
        filename: Union[str, Path],
        num_iteration: Optional[int] = None,
        start_iteration: int = 0,
        importance_type: str = 'split'
    ) -> "Booster":
3930
        """Save Booster to file.
wxchan's avatar
wxchan committed
3931
3932
3933

        Parameters
        ----------
3934
        filename : str or pathlib.Path
3935
            Filename to save Booster.
3936
3937
3938
3939
        num_iteration : int or None, optional (default=None)
            Index of the iteration that should be saved.
            If None, if the best iteration exists, it is saved; otherwise, all iterations are saved.
            If <= 0, all iterations are saved.
Nikita Titov's avatar
Nikita Titov committed
3940
        start_iteration : int, optional (default=0)
3941
            Start index of the iteration that should be saved.
3942
        importance_type : str, optional (default="split")
3943
3944
3945
            What type of feature importance should be saved.
            If "split", result contains numbers of times the feature is used in a model.
            If "gain", result contains total gains of splits which use the feature.
Nikita Titov's avatar
Nikita Titov committed
3946
3947
3948
3949
3950

        Returns
        -------
        self : Booster
            Returns self.
wxchan's avatar
wxchan committed
3951
        """
3952
        if num_iteration is None:
3953
            num_iteration = self.best_iteration
3954
        importance_type_int = _FEATURE_IMPORTANCE_TYPE_MAPPER[importance_type]
wxchan's avatar
wxchan committed
3955
        _safe_call(_LIB.LGBM_BoosterSaveModel(
3956
            self._handle,
3957
            ctypes.c_int(start_iteration),
Guolin Ke's avatar
Guolin Ke committed
3958
            ctypes.c_int(num_iteration),
3959
            ctypes.c_int(importance_type_int),
3960
            _c_str(str(filename))))
3961
        _dump_pandas_categorical(self.pandas_categorical, filename)
Nikita Titov's avatar
Nikita Titov committed
3962
        return self
wxchan's avatar
wxchan committed
3963

3964
3965
3966
3967
3968
    def shuffle_models(
        self,
        start_iteration: int = 0,
        end_iteration: int = -1
    ) -> "Booster":
3969
        """Shuffle models.
Nikita Titov's avatar
Nikita Titov committed
3970

3971
3972
3973
        Parameters
        ----------
        start_iteration : int, optional (default=0)
3974
            The first iteration that will be shuffled.
3975
3976
        end_iteration : int, optional (default=-1)
            The last iteration that will be shuffled.
3977
            If <= 0, means the last available iteration.
3978

Nikita Titov's avatar
Nikita Titov committed
3979
3980
3981
3982
        Returns
        -------
        self : Booster
            Booster with shuffled models.
3983
        """
3984
        _safe_call(_LIB.LGBM_BoosterShuffleModels(
3985
            self._handle,
Guolin Ke's avatar
Guolin Ke committed
3986
3987
            ctypes.c_int(start_iteration),
            ctypes.c_int(end_iteration)))
Nikita Titov's avatar
Nikita Titov committed
3988
        return self
3989

3990
    def model_from_string(self, model_str: str) -> "Booster":
3991
3992
3993
3994
        """Load Booster from a string.

        Parameters
        ----------
3995
        model_str : str
3996
3997
3998
3999
            Model will be loaded from this string.

        Returns
        -------
Nikita Titov's avatar
Nikita Titov committed
4000
        self : Booster
4001
4002
            Loaded Booster object.
        """
4003
4004
4005
        # ensure that existing Booster is freed before replacing it
        # with a new one createdfrom file
        _safe_call(_LIB.LGBM_BoosterFree(self._handle))
4006
        self._free_buffer()
4007
        self._handle = ctypes.c_void_p()
4008
4009
        out_num_iterations = ctypes.c_int(0)
        _safe_call(_LIB.LGBM_BoosterLoadModelFromString(
4010
            _c_str(model_str),
4011
            ctypes.byref(out_num_iterations),
4012
            ctypes.byref(self._handle)))
4013
4014
        out_num_class = ctypes.c_int(0)
        _safe_call(_LIB.LGBM_BoosterGetNumClasses(
4015
            self._handle,
4016
4017
            ctypes.byref(out_num_class)))
        self.__num_class = out_num_class.value
4018
        self.pandas_categorical = _load_pandas_categorical(model_str=model_str)
4019
4020
        return self

4021
4022
4023
4024
4025
4026
    def model_to_string(
        self,
        num_iteration: Optional[int] = None,
        start_iteration: int = 0,
        importance_type: str = 'split'
    ) -> str:
4027
        """Save Booster to string.
4028

4029
4030
4031
4032
4033
4034
        Parameters
        ----------
        num_iteration : int or None, optional (default=None)
            Index of the iteration that should be saved.
            If None, if the best iteration exists, it is saved; otherwise, all iterations are saved.
            If <= 0, all iterations are saved.
Nikita Titov's avatar
Nikita Titov committed
4035
        start_iteration : int, optional (default=0)
4036
            Start index of the iteration that should be saved.
4037
        importance_type : str, optional (default="split")
4038
4039
4040
            What type of feature importance should be saved.
            If "split", result contains numbers of times the feature is used in a model.
            If "gain", result contains total gains of splits which use the feature.
4041
4042
4043

        Returns
        -------
4044
        str_repr : str
4045
4046
            String representation of Booster.
        """
4047
        if num_iteration is None:
4048
            num_iteration = self.best_iteration
4049
        importance_type_int = _FEATURE_IMPORTANCE_TYPE_MAPPER[importance_type]
4050
        buffer_len = 1 << 20
4051
        tmp_out_len = ctypes.c_int64(0)
4052
4053
4054
        string_buffer = ctypes.create_string_buffer(buffer_len)
        ptr_string_buffer = ctypes.c_char_p(*[ctypes.addressof(string_buffer)])
        _safe_call(_LIB.LGBM_BoosterSaveModelToString(
4055
            self._handle,
4056
            ctypes.c_int(start_iteration),
4057
            ctypes.c_int(num_iteration),
4058
            ctypes.c_int(importance_type_int),
4059
            ctypes.c_int64(buffer_len),
4060
4061
4062
            ctypes.byref(tmp_out_len),
            ptr_string_buffer))
        actual_len = tmp_out_len.value
4063
        # if buffer length is not long enough, re-allocate a buffer
4064
4065
4066
4067
        if actual_len > buffer_len:
            string_buffer = ctypes.create_string_buffer(actual_len)
            ptr_string_buffer = ctypes.c_char_p(*[ctypes.addressof(string_buffer)])
            _safe_call(_LIB.LGBM_BoosterSaveModelToString(
4068
                self._handle,
4069
                ctypes.c_int(start_iteration),
4070
                ctypes.c_int(num_iteration),
4071
                ctypes.c_int(importance_type_int),
4072
                ctypes.c_int64(actual_len),
4073
4074
                ctypes.byref(tmp_out_len),
                ptr_string_buffer))
4075
        ret = string_buffer.value.decode('utf-8')
4076
4077
        ret += _dump_pandas_categorical(self.pandas_categorical)
        return ret
4078

4079
4080
4081
4082
4083
4084
4085
    def dump_model(
        self,
        num_iteration: Optional[int] = None,
        start_iteration: int = 0,
        importance_type: str = 'split',
        object_hook: Optional[Callable[[Dict[str, Any]], Dict[str, Any]]] = None
    ) -> Dict[str, Any]:
Nikita Titov's avatar
Nikita Titov committed
4086
        """Dump Booster to JSON format.
wxchan's avatar
wxchan committed
4087

4088
4089
        Parameters
        ----------
4090
4091
4092
4093
        num_iteration : int or None, optional (default=None)
            Index of the iteration that should be dumped.
            If None, if the best iteration exists, it is dumped; otherwise, all iterations are dumped.
            If <= 0, all iterations are dumped.
Nikita Titov's avatar
Nikita Titov committed
4094
        start_iteration : int, optional (default=0)
4095
            Start index of the iteration that should be dumped.
4096
        importance_type : str, optional (default="split")
4097
4098
4099
            What type of feature importance should be dumped.
            If "split", result contains numbers of times the feature is used in a model.
            If "gain", result contains total gains of splits which use the feature.
4100
4101
4102
4103
4104
4105
4106
4107
4108
        object_hook : callable or None, optional (default=None)
            If not None, ``object_hook`` is a function called while parsing the json
            string returned by the C API. It may be used to alter the json, to store
            specific values while building the json structure. It avoids
            walking through the structure again. It saves a significant amount
            of time if the number of trees is huge.
            Signature is ``def object_hook(node: dict) -> dict``.
            None is equivalent to ``lambda node: node``.
            See documentation of ``json.loads()`` for further details.
4109

wxchan's avatar
wxchan committed
4110
4111
        Returns
        -------
4112
        json_repr : dict
Nikita Titov's avatar
Nikita Titov committed
4113
            JSON format of Booster.
wxchan's avatar
wxchan committed
4114
        """
4115
        if num_iteration is None:
4116
            num_iteration = self.best_iteration
4117
        importance_type_int = _FEATURE_IMPORTANCE_TYPE_MAPPER[importance_type]
wxchan's avatar
wxchan committed
4118
        buffer_len = 1 << 20
4119
        tmp_out_len = ctypes.c_int64(0)
wxchan's avatar
wxchan committed
4120
4121
4122
        string_buffer = ctypes.create_string_buffer(buffer_len)
        ptr_string_buffer = ctypes.c_char_p(*[ctypes.addressof(string_buffer)])
        _safe_call(_LIB.LGBM_BoosterDumpModel(
4123
            self._handle,
4124
            ctypes.c_int(start_iteration),
Guolin Ke's avatar
Guolin Ke committed
4125
            ctypes.c_int(num_iteration),
4126
            ctypes.c_int(importance_type_int),
4127
            ctypes.c_int64(buffer_len),
wxchan's avatar
wxchan committed
4128
            ctypes.byref(tmp_out_len),
Guolin Ke's avatar
Guolin Ke committed
4129
            ptr_string_buffer))
wxchan's avatar
wxchan committed
4130
        actual_len = tmp_out_len.value
4131
        # if buffer length is not long enough, reallocate a buffer
wxchan's avatar
wxchan committed
4132
4133
4134
4135
        if actual_len > buffer_len:
            string_buffer = ctypes.create_string_buffer(actual_len)
            ptr_string_buffer = ctypes.c_char_p(*[ctypes.addressof(string_buffer)])
            _safe_call(_LIB.LGBM_BoosterDumpModel(
4136
                self._handle,
4137
                ctypes.c_int(start_iteration),
Guolin Ke's avatar
Guolin Ke committed
4138
                ctypes.c_int(num_iteration),
4139
                ctypes.c_int(importance_type_int),
4140
                ctypes.c_int64(actual_len),
wxchan's avatar
wxchan committed
4141
                ctypes.byref(tmp_out_len),
Guolin Ke's avatar
Guolin Ke committed
4142
                ptr_string_buffer))
4143
        ret = json.loads(string_buffer.value.decode('utf-8'), object_hook=object_hook)
4144
        ret['pandas_categorical'] = json.loads(json.dumps(self.pandas_categorical,
4145
                                                          default=_json_default_with_numpy))
4146
        return ret
wxchan's avatar
wxchan committed
4147

4148
4149
    def predict(
        self,
4150
        data: _LGBM_PredictDataType,
4151
4152
4153
4154
4155
4156
4157
4158
        start_iteration: int = 0,
        num_iteration: Optional[int] = None,
        raw_score: bool = False,
        pred_leaf: bool = False,
        pred_contrib: bool = False,
        data_has_header: bool = False,
        validate_features: bool = False,
        **kwargs: Any
4159
    ) -> Union[np.ndarray, scipy.sparse.spmatrix, List[scipy.sparse.spmatrix]]:
4160
        """Make a prediction.
wxchan's avatar
wxchan committed
4161
4162
4163

        Parameters
        ----------
4164
        data : str, pathlib.Path, numpy array, pandas DataFrame, H2O DataTable's Frame or scipy.sparse
4165
            Data source for prediction.
4166
            If str or pathlib.Path, it represents the path to a text file (CSV, TSV, or LibSVM).
4167
        start_iteration : int, optional (default=0)
4168
            Start index of the iteration to predict.
4169
            If <= 0, starts from the first iteration.
4170
        num_iteration : int or None, optional (default=None)
4171
4172
4173
4174
            Total number of iterations used in the prediction.
            If None, if the best iteration exists and start_iteration <= 0, the best iteration is used;
            otherwise, all iterations from ``start_iteration`` are used (no limits).
            If <= 0, all iterations from ``start_iteration`` are used (no limits).
4175
4176
4177
4178
        raw_score : bool, optional (default=False)
            Whether to predict raw scores.
        pred_leaf : bool, optional (default=False)
            Whether to predict leaf index.
4179
4180
        pred_contrib : bool, optional (default=False)
            Whether to predict feature contributions.
4181

Nikita Titov's avatar
Nikita Titov committed
4182
4183
4184
4185
4186
4187
4188
            .. note::

                If you want to get more explanations for your model's predictions using SHAP values,
                like SHAP interaction values,
                you can install the shap package (https://github.com/slundberg/shap).
                Note that unlike the shap package, with ``pred_contrib`` we return a matrix with an extra
                column, where the last column is the expected value.
4189

4190
4191
        data_has_header : bool, optional (default=False)
            Whether the data has header.
4192
            Used only if data is str.
4193
4194
4195
        validate_features : bool, optional (default=False)
            If True, ensure that the features used to predict match the ones used to train.
            Used only if data is pandas DataFrame.
4196
4197
        **kwargs
            Other parameters for the prediction.
wxchan's avatar
wxchan committed
4198
4199
4200

        Returns
        -------
4201
        result : numpy array, scipy.sparse or list of scipy.sparse
4202
            Prediction result.
4203
            Can be sparse or a list of sparse objects (each element represents predictions for one class) for feature contributions (when ``pred_contrib=True``).
wxchan's avatar
wxchan committed
4204
        """
4205
4206
4207
4208
        predictor = _InnerPredictor.from_booster(
            booster=self,
            pred_parameter=deepcopy(kwargs),
        )
4209
        if num_iteration is None:
4210
            if start_iteration <= 0:
4211
4212
4213
                num_iteration = self.best_iteration
            else:
                num_iteration = -1
4214
4215
4216
4217
4218
4219
4220
4221
4222
4223
        return predictor.predict(
            data=data,
            start_iteration=start_iteration,
            num_iteration=num_iteration,
            raw_score=raw_score,
            pred_leaf=pred_leaf,
            pred_contrib=pred_contrib,
            data_has_header=data_has_header,
            validate_features=validate_features
        )
wxchan's avatar
wxchan committed
4224

4225
4226
    def refit(
        self,
4227
        data: _LGBM_TrainDataType,
4228
        label: _LGBM_LabelType,
4229
4230
        decay_rate: float = 0.9,
        reference: Optional[Dataset] = None,
4231
4232
4233
        weight: Optional[_LGBM_WeightType] = None,
        group: Optional[_LGBM_GroupType] = None,
        init_score: Optional[_LGBM_InitScoreType] = None,
4234
4235
        feature_name: _LGBM_FeatureNameConfiguration = 'auto',
        categorical_feature: _LGBM_CategoricalFeatureConfiguration = 'auto',
4236
4237
4238
        dataset_params: Optional[Dict[str, Any]] = None,
        free_raw_data: bool = True,
        validate_features: bool = False,
4239
        **kwargs
4240
    ) -> "Booster":
Guolin Ke's avatar
Guolin Ke committed
4241
4242
4243
4244
        """Refit the existing Booster by new data.

        Parameters
        ----------
4245
        data : str, pathlib.Path, numpy array, pandas DataFrame, H2O DataTable's Frame, scipy.sparse, Sequence, list of Sequence or list of numpy array
Guolin Ke's avatar
Guolin Ke committed
4246
            Data source for refit.
4247
            If str or pathlib.Path, it represents the path to a text file (CSV, TSV, or LibSVM).
4248
        label : list, numpy 1-D array or pandas Series / one-column DataFrame
Guolin Ke's avatar
Guolin Ke committed
4249
4250
            Label for refit.
        decay_rate : float, optional (default=0.9)
4251
4252
            Decay rate of refit,
            will use ``leaf_output = decay_rate * old_leaf_output + (1.0 - decay_rate) * new_leaf_output`` to refit trees.
4253
4254
        reference : Dataset or None, optional (default=None)
            Reference for ``data``.
4255
4256
4257

            .. versionadded:: 4.0.0

4258
        weight : list, numpy 1-D array, pandas Series or None, optional (default=None)
4259
            Weight for each ``data`` instance. Weights should be non-negative.
4260
4261
4262

            .. versionadded:: 4.0.0

4263
4264
4265
4266
4267
4268
        group : list, numpy 1-D array, pandas Series or None, optional (default=None)
            Group/query size for ``data``.
            Only used in the learning-to-rank task.
            sum(group) = n_samples.
            For example, if you have a 100-document dataset with ``group = [10, 20, 40, 10, 10, 10]``, that means that you have 6 groups,
            where the first 10 records are in the first group, records 11-30 are in the second group, records 31-70 are in the third group, etc.
4269
4270
4271

            .. versionadded:: 4.0.0

4272
4273
        init_score : list, list of lists (for multi-class task), numpy array, pandas Series, pandas DataFrame (for multi-class task), or None, optional (default=None)
            Init score for ``data``.
4274
4275
4276

            .. versionadded:: 4.0.0

4277
4278
4279
        feature_name : list of str, or 'auto', optional (default="auto")
            Feature names for ``data``.
            If 'auto' and data is pandas DataFrame, data columns names are used.
4280
4281
4282

            .. versionadded:: 4.0.0

4283
4284
4285
4286
4287
        categorical_feature : list of str or int, or 'auto', optional (default="auto")
            Categorical features for ``data``.
            If list of int, interpreted as indices.
            If list of str, interpreted as feature names (need to specify ``feature_name`` as well).
            If 'auto' and data is pandas DataFrame, pandas unordered categorical columns are used.
4288
            All values in categorical features will be cast to int32 and thus should be less than int32 max value (2147483647).
4289
4290
4291
            Large values could be memory consuming. Consider using consecutive integers starting from zero.
            All negative values in categorical features will be treated as missing values.
            The output cannot be monotonically constrained with respect to a categorical feature.
4292
            Floating point numbers in categorical features will be rounded towards 0.
4293
4294
4295

            .. versionadded:: 4.0.0

4296
4297
        dataset_params : dict or None, optional (default=None)
            Other parameters for Dataset ``data``.
4298
4299
4300

            .. versionadded:: 4.0.0

4301
4302
        free_raw_data : bool, optional (default=True)
            If True, raw data is freed after constructing inner Dataset for ``data``.
4303
4304
4305

            .. versionadded:: 4.0.0

4306
4307
4308
        validate_features : bool, optional (default=False)
            If True, ensure that the features used to refit the model match the original ones.
            Used only if data is pandas DataFrame.
4309
4310
4311

            .. versionadded:: 4.0.0

4312
4313
        **kwargs
            Other parameters for refit.
4314
            These parameters will be passed to ``predict`` method.
Guolin Ke's avatar
Guolin Ke committed
4315
4316
4317
4318
4319
4320

        Returns
        -------
        result : Booster
            Refitted Booster.
        """
4321
4322
        if self.__set_objective_to_none:
            raise LightGBMError('Cannot refit due to null objective function.')
4323
4324
        if dataset_params is None:
            dataset_params = {}
4325
4326
4327
4328
        predictor = _InnerPredictor.from_booster(
            booster=self,
            pred_parameter=deepcopy(kwargs)
        )
4329
        leaf_preds: np.ndarray = predictor.predict(  # type: ignore[assignment]
4330
4331
4332
4333
4334
            data=data,
            start_iteration=-1,
            pred_leaf=True,
            validate_features=validate_features
        )
4335
        nrow, ncol = leaf_preds.shape
4336
        out_is_linear = ctypes.c_int(0)
4337
        _safe_call(_LIB.LGBM_BoosterGetLinear(
4338
            self._handle,
4339
            ctypes.byref(out_is_linear)))
Nikita Titov's avatar
Nikita Titov committed
4340
4341
4342
4343
4344
        new_params = _choose_param_value(
            main_param_name="linear_tree",
            params=self.params,
            default_value=None
        )
4345
        new_params["linear_tree"] = bool(out_is_linear.value)
4346
4347
4348
4349
4350
4351
4352
4353
4354
4355
4356
4357
4358
        new_params.update(dataset_params)
        train_set = Dataset(
            data=data,
            label=label,
            reference=reference,
            weight=weight,
            group=group,
            init_score=init_score,
            feature_name=feature_name,
            categorical_feature=categorical_feature,
            params=new_params,
            free_raw_data=free_raw_data,
        )
4359
        new_params['refit_decay_rate'] = decay_rate
4360
        new_booster = Booster(new_params, train_set)
Guolin Ke's avatar
Guolin Ke committed
4361
4362
        # Copy models
        _safe_call(_LIB.LGBM_BoosterMerge(
4363
4364
            new_booster._handle,
            predictor._handle))
Guolin Ke's avatar
Guolin Ke committed
4365
        leaf_preds = leaf_preds.reshape(-1)
4366
        ptr_data, _, _ = _c_int_array(leaf_preds)
Guolin Ke's avatar
Guolin Ke committed
4367
        _safe_call(_LIB.LGBM_BoosterRefit(
4368
            new_booster._handle,
Guolin Ke's avatar
Guolin Ke committed
4369
            ptr_data,
4370
4371
            ctypes.c_int32(nrow),
            ctypes.c_int32(ncol)))
4372
        new_booster._network = self._network
Guolin Ke's avatar
Guolin Ke committed
4373
4374
        return new_booster

4375
    def get_leaf_output(self, tree_id: int, leaf_id: int) -> float:
4376
4377
4378
4379
4380
4381
4382
4383
4384
4385
4386
4387
4388
4389
        """Get the output of a leaf.

        Parameters
        ----------
        tree_id : int
            The index of the tree.
        leaf_id : int
            The index of the leaf in the tree.

        Returns
        -------
        result : float
            The output of the leaf.
        """
4390
4391
        ret = ctypes.c_double(0)
        _safe_call(_LIB.LGBM_BoosterGetLeafValue(
4392
            self._handle,
4393
4394
4395
4396
4397
            ctypes.c_int(tree_id),
            ctypes.c_int(leaf_id),
            ctypes.byref(ret)))
        return ret.value

4398
4399
4400
4401
4402
4403
4404
4405
    def set_leaf_output(
        self,
        tree_id: int,
        leaf_id: int,
        value: float,
    ) -> 'Booster':
        """Set the output of a leaf.

4406
4407
        .. versionadded:: 4.0.0

4408
4409
4410
4411
4412
4413
4414
4415
4416
4417
4418
4419
4420
4421
4422
4423
        Parameters
        ----------
        tree_id : int
            The index of the tree.
        leaf_id : int
            The index of the leaf in the tree.
        value : float
            Value to set as the output of the leaf.

        Returns
        -------
        self : Booster
            Booster with the leaf output set.
        """
        _safe_call(
            _LIB.LGBM_BoosterSetLeafValue(
4424
                self._handle,
4425
4426
4427
4428
4429
4430
4431
                ctypes.c_int(tree_id),
                ctypes.c_int(leaf_id),
                ctypes.c_double(value)
            )
        )
        return self

4432
    def num_feature(self) -> int:
4433
4434
4435
4436
4437
4438
4439
        """Get number of features.

        Returns
        -------
        num_feature : int
            The number of features.
        """
4440
4441
        out_num_feature = ctypes.c_int(0)
        _safe_call(_LIB.LGBM_BoosterGetNumFeature(
4442
            self._handle,
4443
4444
4445
            ctypes.byref(out_num_feature)))
        return out_num_feature.value

4446
    def feature_name(self) -> List[str]:
4447
        """Get names of features.
wxchan's avatar
wxchan committed
4448
4449
4450

        Returns
        -------
4451
        result : list of str
4452
            List with names of features.
wxchan's avatar
wxchan committed
4453
        """
4454
        num_feature = self.num_feature()
4455
        # Get name of features
wxchan's avatar
wxchan committed
4456
        tmp_out_len = ctypes.c_int(0)
4457
4458
        reserved_string_buffer_size = 255
        required_string_buffer_size = ctypes.c_size_t(0)
4459
        string_buffers = [ctypes.create_string_buffer(reserved_string_buffer_size) for _ in range(num_feature)]
wxchan's avatar
wxchan committed
4460
4461
        ptr_string_buffers = (ctypes.c_char_p * num_feature)(*map(ctypes.addressof, string_buffers))
        _safe_call(_LIB.LGBM_BoosterGetFeatureNames(
4462
            self._handle,
4463
            ctypes.c_int(num_feature),
wxchan's avatar
wxchan committed
4464
            ctypes.byref(tmp_out_len),
4465
            ctypes.c_size_t(reserved_string_buffer_size),
4466
            ctypes.byref(required_string_buffer_size),
wxchan's avatar
wxchan committed
4467
4468
4469
            ptr_string_buffers))
        if num_feature != tmp_out_len.value:
            raise ValueError("Length of feature names doesn't equal with num_feature")
4470
4471
4472
4473
4474
4475
        actual_string_buffer_size = required_string_buffer_size.value
        # if buffer length is not long enough, reallocate buffers
        if reserved_string_buffer_size < actual_string_buffer_size:
            string_buffers = [ctypes.create_string_buffer(actual_string_buffer_size) for _ in range(num_feature)]
            ptr_string_buffers = (ctypes.c_char_p * num_feature)(*map(ctypes.addressof, string_buffers))
            _safe_call(_LIB.LGBM_BoosterGetFeatureNames(
4476
                self._handle,
4477
4478
4479
4480
4481
                ctypes.c_int(num_feature),
                ctypes.byref(tmp_out_len),
                ctypes.c_size_t(actual_string_buffer_size),
                ctypes.byref(required_string_buffer_size),
                ptr_string_buffers))
4482
        return [string_buffers[i].value.decode('utf-8') for i in range(num_feature)]
wxchan's avatar
wxchan committed
4483

4484
4485
4486
4487
4488
    def feature_importance(
        self,
        importance_type: str = 'split',
        iteration: Optional[int] = None
    ) -> np.ndarray:
4489
        """Get feature importances.
4490

4491
4492
        Parameters
        ----------
4493
        importance_type : str, optional (default="split")
4494
4495
4496
            How the importance is calculated.
            If "split", result contains numbers of times the feature is used in a model.
            If "gain", result contains total gains of splits which use the feature.
4497
4498
4499
4500
        iteration : int or None, optional (default=None)
            Limit number of iterations in the feature importance calculation.
            If None, if the best iteration exists, it is used; otherwise, all trees are used.
            If <= 0, all trees are used (no limits).
4501

4502
4503
        Returns
        -------
4504
4505
        result : numpy array
            Array with feature importances.
4506
        """
4507
4508
        if iteration is None:
            iteration = self.best_iteration
4509
        importance_type_int = _FEATURE_IMPORTANCE_TYPE_MAPPER[importance_type]
4510
        result = np.empty(self.num_feature(), dtype=np.float64)
4511
        _safe_call(_LIB.LGBM_BoosterFeatureImportance(
4512
            self._handle,
4513
4514
4515
            ctypes.c_int(iteration),
            ctypes.c_int(importance_type_int),
            result.ctypes.data_as(ctypes.POINTER(ctypes.c_double))))
4516
        if importance_type_int == _C_API_FEATURE_IMPORTANCE_SPLIT:
4517
            return result.astype(np.int32)
4518
4519
        else:
            return result
4520

4521
4522
4523
4524
4525
4526
    def get_split_value_histogram(
        self,
        feature: Union[int, str],
        bins: Optional[Union[int, str]] = None,
        xgboost_style: bool = False
    ) -> Union[Tuple[np.ndarray, np.ndarray], np.ndarray, pd_DataFrame]:
4527
4528
4529
4530
        """Get split value histogram for the specified feature.

        Parameters
        ----------
4531
        feature : int or str
4532
4533
            The feature name or index the histogram is calculated for.
            If int, interpreted as index.
4534
            If str, interpreted as name.
4535

Nikita Titov's avatar
Nikita Titov committed
4536
4537
4538
            .. warning::

                Categorical features are not supported.
4539

4540
        bins : int, str or None, optional (default=None)
4541
4542
4543
            The maximum number of bins.
            If None, or int and > number of unique split values and ``xgboost_style=True``,
            the number of bins equals number of unique split values.
4544
            If str, it should be one from the list of the supported values by ``numpy.histogram()`` function.
4545
4546
4547
4548
4549
4550
4551
4552
4553
4554
4555
4556
4557
4558
        xgboost_style : bool, optional (default=False)
            Whether the returned result should be in the same form as it is in XGBoost.
            If False, the returned value is tuple of 2 numpy arrays as it is in ``numpy.histogram()`` function.
            If True, the returned value is matrix, in which the first column is the right edges of non-empty bins
            and the second one is the histogram values.

        Returns
        -------
        result_tuple : tuple of 2 numpy arrays
            If ``xgboost_style=False``, the values of the histogram of used splitting values for the specified feature
            and the bin edges.
        result_array_like : numpy array or pandas DataFrame (if pandas is installed)
            If ``xgboost_style=True``, the histogram of used splitting values for the specified feature.
        """
4559
        def add(root: Dict[str, Any]) -> None:
4560
4561
            """Recursively add thresholds."""
            if 'split_index' in root:  # non-leaf
4562
                if feature_names is not None and isinstance(feature, str):
4563
4564
4565
4566
                    split_feature = feature_names[root['split_feature']]
                else:
                    split_feature = root['split_feature']
                if split_feature == feature:
4567
                    if isinstance(root['threshold'], str):
4568
4569
4570
                        raise LightGBMError('Cannot compute split value histogram for the categorical feature')
                    else:
                        values.append(root['threshold'])
4571
4572
4573
4574
4575
4576
                add(root['left_child'])
                add(root['right_child'])

        model = self.dump_model()
        feature_names = model.get('feature_names')
        tree_infos = model['tree_info']
4577
        values: List[float] = []
4578
4579
4580
        for tree_info in tree_infos:
            add(tree_info['tree_structure'])

4581
        if bins is None or isinstance(bins, int) and xgboost_style:
4582
4583
4584
4585
4586
4587
4588
            n_unique = len(np.unique(values))
            bins = max(min(n_unique, bins) if bins is not None else n_unique, 1)
        hist, bin_edges = np.histogram(values, bins=bins)
        if xgboost_style:
            ret = np.column_stack((bin_edges[1:], hist))
            ret = ret[ret[:, 1] > 0]
            if PANDAS_INSTALLED:
4589
                return pd_DataFrame(ret, columns=['SplitValue', 'Count'])
4590
4591
4592
4593
4594
            else:
                return ret
        else:
            return hist, bin_edges

4595
4596
4597
4598
    def __inner_eval(
        self,
        data_name: str,
        data_idx: int,
4599
        feval: Optional[Union[_LGBM_CustomEvalFunction, List[_LGBM_CustomEvalFunction]]]
4600
    ) -> List[_LGBM_BoosterEvalMethodResultType]:
4601
        """Evaluate training or validation data."""
wxchan's avatar
wxchan committed
4602
        if data_idx >= self.__num_dataset:
4603
            raise ValueError("Data_idx should be smaller than number of dataset")
wxchan's avatar
wxchan committed
4604
4605
4606
        self.__get_eval_info()
        ret = []
        if self.__num_inner_eval > 0:
4607
            result = np.empty(self.__num_inner_eval, dtype=np.float64)
Guolin Ke's avatar
Guolin Ke committed
4608
            tmp_out_len = ctypes.c_int(0)
wxchan's avatar
wxchan committed
4609
            _safe_call(_LIB.LGBM_BoosterGetEval(
4610
                self._handle,
Guolin Ke's avatar
Guolin Ke committed
4611
                ctypes.c_int(data_idx),
wxchan's avatar
wxchan committed
4612
                ctypes.byref(tmp_out_len),
Guolin Ke's avatar
Guolin Ke committed
4613
                result.ctypes.data_as(ctypes.POINTER(ctypes.c_double))))
wxchan's avatar
wxchan committed
4614
            if tmp_out_len.value != self.__num_inner_eval:
4615
                raise ValueError("Wrong length of eval results")
4616
            for i in range(self.__num_inner_eval):
4617
4618
                ret.append((data_name, self.__name_inner_eval[i],
                            result[i], self.__higher_better_inner_eval[i]))
4619
4620
        if callable(feval):
            feval = [feval]
wxchan's avatar
wxchan committed
4621
4622
4623
4624
4625
        if feval is not None:
            if data_idx == 0:
                cur_data = self.train_set
            else:
                cur_data = self.valid_sets[data_idx - 1]
4626
4627
4628
4629
4630
4631
4632
4633
4634
            for eval_function in feval:
                if eval_function is None:
                    continue
                feval_ret = eval_function(self.__inner_predict(data_idx), cur_data)
                if isinstance(feval_ret, list):
                    for eval_name, val, is_higher_better in feval_ret:
                        ret.append((data_name, eval_name, val, is_higher_better))
                else:
                    eval_name, val, is_higher_better = feval_ret
wxchan's avatar
wxchan committed
4635
4636
4637
                    ret.append((data_name, eval_name, val, is_higher_better))
        return ret

4638
    def __inner_predict(self, data_idx: int) -> np.ndarray:
4639
        """Predict for training and validation dataset."""
wxchan's avatar
wxchan committed
4640
        if data_idx >= self.__num_dataset:
4641
            raise ValueError("Data_idx should be smaller than number of dataset")
wxchan's avatar
wxchan committed
4642
4643
4644
4645
4646
        if self.__inner_predict_buffer[data_idx] is None:
            if data_idx == 0:
                n_preds = self.train_set.num_data() * self.__num_class
            else:
                n_preds = self.valid_sets[data_idx - 1].num_data() * self.__num_class
4647
            self.__inner_predict_buffer[data_idx] = np.empty(n_preds, dtype=np.float64)
4648
        # avoid to predict many time in one iteration
wxchan's avatar
wxchan committed
4649
4650
        if not self.__is_predicted_cur_iter[data_idx]:
            tmp_out_len = ctypes.c_int64(0)
4651
            data_ptr = self.__inner_predict_buffer[data_idx].ctypes.data_as(ctypes.POINTER(ctypes.c_double))  # type: ignore[union-attr]
wxchan's avatar
wxchan committed
4652
            _safe_call(_LIB.LGBM_BoosterGetPredict(
4653
                self._handle,
Guolin Ke's avatar
Guolin Ke committed
4654
                ctypes.c_int(data_idx),
wxchan's avatar
wxchan committed
4655
4656
                ctypes.byref(tmp_out_len),
                data_ptr))
4657
            if tmp_out_len.value != len(self.__inner_predict_buffer[data_idx]):  # type: ignore[arg-type]
4658
                raise ValueError(f"Wrong length of predict results for data {data_idx}")
wxchan's avatar
wxchan committed
4659
            self.__is_predicted_cur_iter[data_idx] = True
4660
        result: np.ndarray = self.__inner_predict_buffer[data_idx]  # type: ignore[assignment]
4661
4662
4663
4664
        if self.__num_class > 1:
            num_data = result.size // self.__num_class
            result = result.reshape(num_data, self.__num_class, order='F')
        return result
wxchan's avatar
wxchan committed
4665

4666
    def __get_eval_info(self) -> None:
4667
        """Get inner evaluation count and names."""
wxchan's avatar
wxchan committed
4668
4669
        if self.__need_reload_eval_info:
            self.__need_reload_eval_info = False
Guolin Ke's avatar
Guolin Ke committed
4670
            out_num_eval = ctypes.c_int(0)
4671
            # Get num of inner evals
wxchan's avatar
wxchan committed
4672
            _safe_call(_LIB.LGBM_BoosterGetEvalCounts(
4673
                self._handle,
wxchan's avatar
wxchan committed
4674
4675
4676
                ctypes.byref(out_num_eval)))
            self.__num_inner_eval = out_num_eval.value
            if self.__num_inner_eval > 0:
4677
                # Get name of eval metrics
Guolin Ke's avatar
Guolin Ke committed
4678
                tmp_out_len = ctypes.c_int(0)
4679
4680
4681
                reserved_string_buffer_size = 255
                required_string_buffer_size = ctypes.c_size_t(0)
                string_buffers = [
4682
                    ctypes.create_string_buffer(reserved_string_buffer_size) for _ in range(self.__num_inner_eval)
4683
                ]
wxchan's avatar
wxchan committed
4684
                ptr_string_buffers = (ctypes.c_char_p * self.__num_inner_eval)(*map(ctypes.addressof, string_buffers))
wxchan's avatar
wxchan committed
4685
                _safe_call(_LIB.LGBM_BoosterGetEvalNames(
4686
                    self._handle,
4687
                    ctypes.c_int(self.__num_inner_eval),
wxchan's avatar
wxchan committed
4688
                    ctypes.byref(tmp_out_len),
4689
                    ctypes.c_size_t(reserved_string_buffer_size),
4690
                    ctypes.byref(required_string_buffer_size),
wxchan's avatar
wxchan committed
4691
4692
                    ptr_string_buffers))
                if self.__num_inner_eval != tmp_out_len.value:
4693
                    raise ValueError("Length of eval names doesn't equal with num_evals")
4694
4695
4696
4697
4698
4699
4700
4701
                actual_string_buffer_size = required_string_buffer_size.value
                # if buffer length is not long enough, reallocate buffers
                if reserved_string_buffer_size < actual_string_buffer_size:
                    string_buffers = [
                        ctypes.create_string_buffer(actual_string_buffer_size) for _ in range(self.__num_inner_eval)
                    ]
                    ptr_string_buffers = (ctypes.c_char_p * self.__num_inner_eval)(*map(ctypes.addressof, string_buffers))
                    _safe_call(_LIB.LGBM_BoosterGetEvalNames(
4702
                        self._handle,
4703
4704
4705
4706
4707
4708
4709
4710
4711
4712
4713
                        ctypes.c_int(self.__num_inner_eval),
                        ctypes.byref(tmp_out_len),
                        ctypes.c_size_t(actual_string_buffer_size),
                        ctypes.byref(required_string_buffer_size),
                        ptr_string_buffers))
                self.__name_inner_eval = [
                    string_buffers[i].value.decode('utf-8') for i in range(self.__num_inner_eval)
                ]
                self.__higher_better_inner_eval = [
                    name.startswith(('auc', 'ndcg@', 'map@', 'average_precision')) for name in self.__name_inner_eval
                ]