change to rel import within dgl

python/dgl/backend/__init__.py

+from __future__ import absolute_import
+
 import os
+
 __backend__ = os.environ.get('DGLBACKEND', 'pytorch').lower()
 if __backend__ == 'numpy':
-    from dgl.backend.numpy import *
+    from .numpy import *
 elif __backend__ == 'pytorch':
-    from dgl.backend.pytorch import *
+    from .pytorch import *
 else:
     raise Exception("Unsupported backend %s" % __backend__)

python/dgl/backend/pytorch.py

@@ -4,7 +4,7 @@ import torch as th
 
 from .._ffi.runtime_ctypes import TVMType, TVMContext, TVMArray
 from .._ffi.runtime_ctypes import TypeCode, tvm_shape_index_t
-from ..context as cpu, gpu
+from ..context import cpu, gpu
 
 # Tensor types
 Tensor = th.Tensor

python/dgl/batch.py

@@ -3,9 +3,8 @@ from __future__ import absolute_import
 
 import numpy as np
 
-from dgl.graph import DGLGraph
-import dgl.backend as F
-import dgl
+from .graph import DGLGraph
+from . import backend as F
 
 class BatchedDGLGraph(DGLGraph):
     def __init__(self, graph_list, node_attrs=None, edge_attrs=None, **attr):

python/dgl/cached_graph.py

@@ -7,9 +7,9 @@ from __future__ import absolute_import
 
 import igraph
 
-import dgl.backend as F
-from dgl.backend import Tensor
-import dgl.utils as utils
+from . import backend as F
+from .backend import Tensor
+from . import utils
 
 class CachedGraph:
     def __init__(self):

python/dgl/cgraph.py

 from __future__ import absolute_import
 
 from ._ffi.function import _init_api
-import .backend as F
+from . import backend as F
 
 class DGLGraph(object):
     def __init__(self):

python/dgl/data/citation_graph.py

@@ -11,7 +11,7 @@ import networkx as nx
 import scipy.sparse as sp
 import os, sys
 
-from dgl.data.utils import download, extract_archive, get_download_dir
+from .utils import download, extract_archive, get_download_dir
 
 _urls = {
     'cora' : 'https://www.dropbox.com/s/3ggdpkj7ou8svoc/cora.zip?dl=1',

python/dgl/data/tree.py

@@ -10,9 +10,9 @@ from nltk.tree import Tree
 from nltk.corpus.reader import BracketParseCorpusReader
 import networkx as nx
 
-import dgl
-import dgl.backend as F
-from dgl.data.utils import download, extract_archive, get_download_dir
+from .. import backend as F
+from ..graph import DGLGraph
+from .utils import download, extract_archive, get_download_dir
 
 _urls = {
     'sst' : 'https://www.dropbox.com/s/dw8kr2vuq7k4dqi/sst.zip?dl=1',

python/dgl/frame.py

@@ -4,9 +4,9 @@ from __future__ import absolute_import
 from collections import MutableMapping
 import numpy as np
 
-import dgl.backend as F
-from dgl.backend import Tensor
-import dgl.utils as utils
+from . import backend as F
+from .backend import Tensor
+from . import utils
 
 class Frame(MutableMapping):
     def __init__(self, data=None):

python/dgl/function/__init__.py

-from .message import *
+"""DGL builtin functors"""
+from __future__ import absolute_import
 
+from .message import *
 from .reducer import *

python/dgl/function/reducer.py

 """Built-in reducer function."""
 from __future__ import absolute_import
 
-import dgl.backend as F
+from .. import backend as F
 
 __all__ = ["ReduceFunction", "sum", "max"]
 

python/dgl/generator/__init__.py

+"""Package for graph generators"""
+from __future__ import absolute_import
+
 from .line import *

python/dgl/generator/line.py

@@ -4,9 +4,9 @@ from __future__ import absolute_import
 import networkx as nx
 import numpy as np
 
-import dgl.backend as F
-from dgl.graph import DGLGraph
-from dgl.frame import FrameRef
+from .. import backend as F
+from ..graph import DGLGraph
+from ..frame import FrameRef
 
 def line_graph(G, no_backtracking=False):
     """Create the line graph that shares the underlying features.

python/dgl/graph.py

@@ -6,15 +6,15 @@ import networkx as nx
 from networkx.classes.digraph import DiGraph
 
 import dgl
-from dgl.base import ALL, is_all, __MSG__, __REPR__
-import dgl.backend as F
-from dgl.backend import Tensor
-from dgl.cached_graph import CachedGraph, create_cached_graph
-import dgl.context as context
-from dgl.frame import FrameRef, merge_frames
-from dgl.nx_adapt import nx_init
-import dgl.scheduler as scheduler
-import dgl.utils as utils
+from .base import ALL, is_all, __MSG__, __REPR__
+from . import backend as F
+from .backend import Tensor
+from .cached_graph import CachedGraph, create_cached_graph
+from . import context
+from .frame import FrameRef, merge_frames
+from .nx_adapt import nx_init
+from . import scheduler
+from . import utils
 
 class DGLGraph(DiGraph):
     """Base graph class specialized for neural networks on graphs.

python/dgl/nn/__init__.py

+"""Package nn modules"""
+from __future__ import absolute_import
+
 import os
+
 __backend__ = os.environ.get('DGLBACKEND', 'pytorch').lower()
 
 if __backend__ == 'numpy':

python/dgl/nn/pytorch/gcn.py

@@ -7,9 +7,8 @@ GCN with SPMV specialization.
 """
 import torch.nn as nn
 
-import dgl
-import dgl.function as fn
-from dgl.base import ALL, is_all
+from ... import function as fn
+from ...base import ALL, is_all
 
 class NodeUpdateModule(nn.Module):
     def __init__(self, in_feats, out_feats, activation=None):

python/dgl/scheduler.py

@@ -3,10 +3,10 @@ from __future__ import absolute_import
 
 import numpy as np
 
-import dgl.backend as F
-import dgl.function.message as fmsg
-import dgl.function.reducer as fred
-import dgl.utils as utils
+from . import backend as F
+from .function import message as fmsg
+from .function import reducer as fred
+from . import utils
 
 __all__ = ["degree_bucketing", "get_executor"]
 

python/dgl/subgraph.py

@@ -2,11 +2,12 @@
 from __future__ import absolute_import
 
 import networkx as nx
-import dgl.backend as F
-from dgl.frame import Frame, FrameRef
-from dgl.graph import DGLGraph
-from dgl.nx_adapt import nx_init
-import dgl.utils as utils
+
+from . import backend as F
+from .frame import Frame, FrameRef
+from .graph import DGLGraph
+from .nx_adapt import nx_init
+from . import utils
 
 class DGLSubGraph(DGLGraph):
     # TODO(gaiyu): ReadOnlyGraph

python/dgl/utils.py

@@ -5,8 +5,8 @@ from collections import Mapping
 from functools import wraps
 import numpy as np
 
-import dgl.backend as F
-from dgl.backend import Tensor, SparseTensor
+from . import backend as F
+from .backend import Tensor, SparseTensor
 
 def is_id_tensor(u):
     """Return whether the input is a supported id tensor."""