GitLab

* python-package: fix creating eval_set in LGBMClassifier

* replace elements in eval_set directly

* change whitespace to underline in feature names

* add test

* fix bug

* fix bug

* warning -> fatal

* python-package: add graphviz.Digraph parameters

* examples: add a plottig example with graphviz

* fix tree index in print

* python-package: refine plot_tree

This change splits plot_tree to two methods:
1. method of creating digraph of tree
2. method of plotting the digraph with matplotlib

* fix doc

* fix doc

* Python: Fix RandomState issue #376

* Add test case for Python's Shuffle=True

np.random.seed(seed) in _make_n_folds(...) has a side effect setting random seed for outer scope

* add data_splitter for cv

* update gitignore

* clean code

When setting the parameter 'shuffle' of cv method to False ,the program raises the error below:
UnboundLocalError:local variable 'randidx' referenced before assignment

fixed.

* refine plot

* use warnings

* refine  logic

* revert 'move to compat.py'

* Support Poisson regression

Close #233

* update docs

* add plot metrics

* move 'raise Exception' to check_not_tuple_of_2_elements

* rename 'plot_metrics' to 'plot_metric'

* fix misleading message/docs

* change 'Metrics' in title to 'Metric'

* fix misleading comment

* add plot tree

* add docs

* add example

* add test

* fix test

* fix decision type

* add show_info

* use feature name if available

* use json instead of repr/eval for pandas_categorical

* fix json dumps with numpy data

* add more test cases

* expose save model to string to c_api

* add 'save-model-to-string' to python; use it to copy model

* remove boosting_type

* add plot importance

* add plot example

* fix bug for categorical_feature

* add test on load model with categorical feature

* add unseen category in test dataset

* save/load pandas_categorical to model

* fix logic

* cast pandas columns to string

* add load pandas_categorical from file to _InnerPredictor init

* suppprt pandas categorical

* refine logic

* make default=auto

* fix train/valid categorical codes

* add test

* unify set _predictor

* fix tests

* fix warning

* support feature_name=int

* add compat.py

* unify types

* range/xrange -> range_

* move argc_ to compat.py; add is_numeric error type

* use simplejson for json

* move json to compat.py

* move pandas to compat.py

* move sklearn to compat.py

* remove unused function

* fix 'unify types'

* argc_ (lambda -> def)