[DGL-Go][Doc] Update DGL-Go version to 0.0.2 and misc fix from bug ba…

…sh (dmlc#4236) * Update * Update * Update * Update Co-authored-by: Ubuntu <[email protected]> Co-authored-by: Xin Yao <[email protected]>
ColaBH · Jul 14, 2022 · fdbf5a0 · fdbf5a0
1 parent 79b0a50
commit fdbf5a0
Show file tree

Hide file tree

Showing 29 changed files with 45 additions and 48 deletions.
diff --git a/dglgo/README.md b/dglgo/README.md
@@ -61,7 +61,7 @@ Let's use one of the most classical setups -- training a GraphSAGE model for nod
 classification on the Cora citation graph dataset as an
 example.
 
-### Step one: `dgl configure`
+### Step 1: `dgl configure`
 
 First step, use `dgl configure` to generate a YAML configuration file.
 
@@ -85,7 +85,7 @@ At this point you can also change options to explore optimization potentials.
 The snippet below shows the configuration file generated by the command above.
 
 ```yaml
-version: 0.0.1
+version: 0.0.2
 pipeline_name: nodepred
 pipeline_mode: train
 device: cpu
@@ -181,7 +181,7 @@ That's all! Basically you only need two commands to train a graph neural network
 
 ### Step 3: `dgl export` for more advanced customization
 
-That's not everything yet. You may want to open the hood and and invoke deeper
+That's not everything yet. You may want to open the hood and invoke deeper
 customization. DGL-Go can export a **self-contained, reproducible** Python
 script for you to do anything you like.
 

diff --git a/dglgo/dglgo/utils/enter_config.py b/dglgo/dglgo/utils/enter_config.py
@@ -23,7 +23,7 @@ class PipelineConfig(DGLBaseModel):
     loss: str = "CrossEntropyLoss"
 
 class UserConfig(DGLBaseModel):
-    version: Optional[str] = "0.0.1"
+    version: Optional[str] = "0.0.2"
     pipeline_name: PipelineFactory.get_pipeline_enum()
     pipeline_mode: str
     device: str = "cpu"
diff --git a/dglgo/recipes/graphpred_hiv_gin.yaml b/dglgo/recipes/graphpred_hiv_gin.yaml
@@ -1,4 +1,4 @@
-version: 0.0.1
+version: 0.0.2
 pipeline_name: graphpred
 pipeline_mode: train
 device: cuda:0                # Torch device name, e.q. cpu or cuda or cuda:0

diff --git a/dglgo/recipes/graphpred_hiv_pna.yaml b/dglgo/recipes/graphpred_hiv_pna.yaml
@@ -1,4 +1,4 @@
-version: 0.0.1
+version: 0.0.2
 pipeline_name: graphpred
 pipeline_mode: train
 device: cuda:0                # Torch device name, e.q. cpu or cuda or cuda:0

diff --git a/dglgo/recipes/graphpred_pcba_gin.yaml b/dglgo/recipes/graphpred_pcba_gin.yaml
@@ -1,4 +1,4 @@
-version: 0.0.1
+version: 0.0.2
 pipeline_name: graphpred
 pipeline_mode: train
 device: cuda:0                # Torch device name, e.q. cpu or cuda or cuda:0

diff --git a/dglgo/recipes/linkpred_citation2_sage.yaml b/dglgo/recipes/linkpred_citation2_sage.yaml
@@ -1,4 +1,4 @@
-version: 0.0.1
+version: 0.0.2
 pipeline_name: linkpred
 pipeline_mode: train
 device: cpu

diff --git a/dglgo/recipes/linkpred_collab_sage.yaml b/dglgo/recipes/linkpred_collab_sage.yaml
@@ -1,4 +1,4 @@
-version: 0.0.1
+version: 0.0.2
 pipeline_name: linkpred
 pipeline_mode: train
 device: cpu

diff --git a/dglgo/recipes/linkpred_cora_sage.yaml b/dglgo/recipes/linkpred_cora_sage.yaml
@@ -1,4 +1,4 @@
-version: 0.0.1
+version: 0.0.2
 pipeline_name: linkpred
 pipeline_mode: train
 device: cuda

diff --git a/dglgo/recipes/nodepred-ns_arxiv_gcn.yaml b/dglgo/recipes/nodepred-ns_arxiv_gcn.yaml
@@ -1,5 +1,5 @@
 # Accuracy across 5 runs: 0.593288 ± 0.006103
-version: 0.0.1
+version: 0.0.2
 pipeline_name: nodepred-ns
 pipeline_mode: train
 device: 'cuda:0'

diff --git a/dglgo/recipes/nodepred-ns_product_sage.yaml b/dglgo/recipes/nodepred-ns_product_sage.yaml
@@ -1,5 +1,5 @@
 # Accuracy across 1 runs: 0.796911
-version: 0.0.1
+version: 0.0.2
 pipeline_name: nodepred-ns
 pipeline_mode: train
 device: cuda

diff --git a/dglgo/recipes/nodepred_citeseer_gat.yaml b/dglgo/recipes/nodepred_citeseer_gat.yaml
@@ -1,5 +1,5 @@
 # Accuracy across 10 runs: 0.7097 ± 0.006914
-version: 0.0.1
+version: 0.0.2
 pipeline_name: nodepred
 pipeline_mode: train
 device: cuda:0

diff --git a/dglgo/recipes/nodepred_citeseer_gcn.yaml b/dglgo/recipes/nodepred_citeseer_gcn.yaml
@@ -1,5 +1,5 @@
 # Accuracy across 10 runs: 0.6852 ± 0.008875
-version: 0.0.1
+version: 0.0.2
 pipeline_name: nodepred
 pipeline_mode: train
 device: cuda:0

diff --git a/dglgo/recipes/nodepred_citeseer_sage.yaml b/dglgo/recipes/nodepred_citeseer_sage.yaml
@@ -1,5 +1,5 @@
 # Accuracy across 10 runs: 0.6994 ± 0.004005
-version: 0.0.1
+version: 0.0.2
 pipeline_name: nodepred
 pipeline_mode: train
 device: cuda:0

diff --git a/dglgo/recipes/nodepred_cora_gat.yaml b/dglgo/recipes/nodepred_cora_gat.yaml
@@ -1,5 +1,5 @@
 # Accuracy across 10 runs: 0.8208 ± 0.00663
-version: 0.0.1
+version: 0.0.2
 pipeline_name: nodepred
 pipeline_mode: train
 device: cuda:0

diff --git a/dglgo/recipes/nodepred_cora_gcn.yaml b/dglgo/recipes/nodepred_cora_gcn.yaml
@@ -1,5 +1,5 @@
 # Accuracy across 10 runs: 0.802 ± 0.005329
-version: 0.0.1
+version: 0.0.2
 pipeline_name: nodepred
 pipeline_mode: train
 device: cuda:0

diff --git a/dglgo/recipes/nodepred_cora_sage.yaml b/dglgo/recipes/nodepred_cora_sage.yaml
@@ -1,5 +1,5 @@
 # Accuracy across 10 runs: 0.8163 ± 0.006856
-version: 0.0.1
+version: 0.0.2
 pipeline_name: nodepred
 pipeline_mode: train
 device: cuda:0

diff --git a/dglgo/recipes/nodepred_pubmed_gat.yaml b/dglgo/recipes/nodepred_pubmed_gat.yaml
@@ -1,5 +1,5 @@
 # Accuracy across 10 runs: 0.7788 ± 0.002227
-version: 0.0.1
+version: 0.0.2
 pipeline_name: nodepred
 pipeline_mode: train
 device: cuda:0

diff --git a/dglgo/recipes/nodepred_pubmed_gcn.yaml b/dglgo/recipes/nodepred_pubmed_gcn.yaml
@@ -1,5 +1,5 @@
 # Accuracy across 10 runs: 0.7826 ± 0.004317
-version: 0.0.1
+version: 0.0.2
 pipeline_name: nodepred
 pipeline_mode: train
 device: cuda:0

diff --git a/dglgo/recipes/nodepred_pubmed_sage.yaml b/dglgo/recipes/nodepred_pubmed_sage.yaml
@@ -1,5 +1,5 @@
 # Accuracy across 10 runs: 0.7819 ± 0.003176
-version: 0.0.1
+version: 0.0.2
 pipeline_name: nodepred
 pipeline_mode: train
 device: cuda:0

diff --git a/dglgo/setup.py b/dglgo/setup.py
@@ -4,7 +4,7 @@
 from distutils.core import setup
 
 setup(name='dglgo',
-      version='0.0.1',
+      version='0.0.2',
       description='DGL',
       author='DGL Team',
       author_email='[email protected]',

diff --git a/dglgo/tests/cfg.yml b/dglgo/tests/cfg.yml
@@ -1,4 +1,4 @@
-version: 0.0.1
+version: 0.0.2
 pipeline_name: nodepred
 pipeline_mode: train
 device: cpu

diff --git a/python/dgl/data/citation_graph.py b/python/dgl/data/citation_graph.py
@@ -439,7 +439,7 @@ def __getitem__(self, idx):
 
             graph structure, node features and labels.
 
-            - ``ndata['train_mask']``： mask for training node set
+            - ``ndata['train_mask']``: mask for training node set
             - ``ndata['val_mask']``: mask for validation node set
             - ``ndata['test_mask']``: mask for test node set
             - ``ndata['feat']``: node feature
@@ -590,7 +590,7 @@ def __getitem__(self, idx):
 
             graph structure, node features and labels.
 
-            - ``ndata['train_mask']``： mask for training node set
+            - ``ndata['train_mask']``: mask for training node set
             - ``ndata['val_mask']``: mask for validation node set
             - ``ndata['test_mask']``: mask for test node set
             - ``ndata['feat']``: node feature
@@ -738,7 +738,7 @@ def __getitem__(self, idx):
 
             graph structure, node features and labels.
 
-            - ``ndata['train_mask']``： mask for training node set
+            - ``ndata['train_mask']``: mask for training node set
             - ``ndata['val_mask']``: mask for validation node set
             - ``ndata['test_mask']``: mask for test node set
             - ``ndata['feat']``: node feature

diff --git a/python/dgl/data/dgl_dataset.py b/python/dgl/data/dgl_dataset.py
@@ -8,7 +8,6 @@
 import abc
 from .utils import download, extract_archive, get_download_dir, makedirs
 from ..utils import retry_method_with_fix
-from .._ffi.base import __version__
 
 class DGLDataset(object):
     r"""The basic DGL dataset for creating graph datasets.
@@ -238,17 +237,13 @@ def raw_path(self):
     def save_dir(self):
         r"""Directory to save the processed dataset.
         """
-        return self._save_dir + "_v{}".format(__version__)
+        return self._save_dir
 
     @property
     def save_path(self):
         r"""Path to save the processed dataset.
         """
-        if hasattr(self, '_reorder'):
-            path = 'reordered' if self._reorder else 'un_reordered'
-            return os.path.join(self._save_dir, self.name, path)
-        else:
-            return os.path.join(self._save_dir, self.name)
+        return os.path.join(self._save_dir, self.name)
 
     @property
     def verbose(self):

diff --git a/python/dgl/data/flickr.py b/python/dgl/data/flickr.py
@@ -50,6 +50,7 @@ class FlickrDataset(DGLBuiltinDataset):
 
     Examples
     --------
+    >>> from dgl.data import FlickrDataset
     >>> dataset = FlickrDataset()
     >>> dataset.num_classes
     7
@@ -151,9 +152,9 @@ def __getitem__(self, idx):
 
             - ``ndata['label']``: node label
             - ``ndata['feat']``: node feature
-            - ``ndata['train_mask']``： mask for training node set
+            - ``ndata['train_mask']``: mask for training node set
             - ``ndata['val_mask']``: mask for validation node set
-            - ``ndata['test_mask']:`` mask for test node set
+            - ``ndata['test_mask']``: mask for test node set
 
         """
         assert idx == 0, "This dataset has only one graph"

diff --git a/python/dgl/data/utils.py b/python/dgl/data/utils.py
@@ -17,7 +17,6 @@
 from .tensor_serialize import save_tensors, load_tensors
 
 from .. import backend as F
-from .._ffi.base import __version__
 
 __all__ = ['loadtxt','download', 'check_sha1', 'extract_archive',
         'get_download_dir', 'Subset', 'split_dataset', 'save_graphs',
@@ -241,7 +240,7 @@ def get_download_dir():
     dirname : str
         Path to the download directory
     """
-    default_dir = os.path.join(os.path.expanduser('~'), '.dgl_v{}'.format(__version__))
+    default_dir = os.path.join(os.path.expanduser('~'), '.dgl')
     dirname = os.environ.get('DGL_DOWNLOAD_DIR', default_dir)
     if not os.path.exists(dirname):
         os.makedirs(dirname)

diff --git a/python/dgl/data/wikics.py b/python/dgl/data/wikics.py
@@ -50,6 +50,7 @@ class WikiCSDataset(DGLBuiltinDataset):
 
     Examples
     --------
+    >>> from dgl.data import WikiCSDataset
     >>> dataset = WikiCSDataset()
     >>> dataset.num_classes
     10

diff --git a/python/dgl/data/yelp.py b/python/dgl/data/yelp.py
@@ -151,9 +151,9 @@ def __getitem__(self, idx):
 
             - ``ndata['label']``: node label
             - ``ndata['feat']``: node feature
-            - ``ndata['train_mask']``： mask for training node set
+            - ``ndata['train_mask']``: mask for training node set
             - ``ndata['val_mask']``: mask for validation node set
-            - ``ndata['test_mask']:`` mask for test node set
+            - ``ndata['test_mask']``: mask for test node set
 
         """
         assert idx == 0, "This dataset has only one graph"

diff --git a/python/dgl/nn/pytorch/conv/egatconv.py b/python/dgl/nn/pytorch/conv/egatconv.py
@@ -45,7 +45,7 @@ class EGATConv(nn.Module):
     num_heads : int
         Number of attention heads.
     bias : bool, optional
-        If True, add bias term to :math: `f_{ij}^{\prime}`. Defaults: ``True``.
+        If True, add bias term to :math:`f_{ij}^{\prime}`. Defaults: ``True``.
 
     Examples
     ----------
@@ -170,16 +170,16 @@ def forward(self, graph, nfeats, efeats, get_attention=False):
         Returns
         -------
         pair of torch.Tensor
-            node output features followed by edge output features
-            The node output feature of shape :math:`(N, H, D_{out})`
-            The edge output feature of shape :math:`(F, H, F_{out})`
+            node output features followed by edge output features.
+            The node output feature is of shape :math:`(N, H, D_{out})`
+            The edge output feature is of shape :math:`(F, H, F_{out})`
             where:
                 :math:`H` is the number of heads,
                 :math:`D_{out}` is size of output node feature,
                 :math:`F_{out}` is size of output edge feature.
         torch.Tensor, optional
             The attention values of shape :math:`(E, H, 1)`.
-            This is returned only when :attr: `get_attention` is ``True``.
+            This is returned only when :attr:`get_attention` is ``True``.
         """
 
         with graph.local_scope():

diff --git a/python/dgl/transforms/functional.py b/python/dgl/transforms/functional.py
@@ -2872,6 +2872,8 @@ def sort_csr_by_tag(g, tag, tag_offset_name='_TAG_OFFSET', tag_type='node'):
     ``tag_type`` is ``node``.
 
     >>> import dgl
+    >>> import torch
+
     >>> g = dgl.graph(([0,0,0,0,0,1,1,1],[0,1,2,3,4,0,1,2]))
     >>> g.adjacency_matrix(scipy_fmt='csr').nonzero()
     (array([0, 0, 0, 0, 0, 1, 1, 1], dtype=int32),
@@ -2890,11 +2892,10 @@ def sort_csr_by_tag(g, tag, tag_offset_name='_TAG_OFFSET', tag_type='node'):
 
     ``tag_type`` is ``edge``.
 
-    >>> from dgl import backend as F
     >>> g = dgl.graph(([0,0,0,0,0,1,1,1],[0,1,2,3,4,0,1,2]))
     >>> g.edges()
     (tensor([0, 0, 0, 0, 0, 1, 1, 1]), tensor([0, 1, 2, 3, 4, 0, 1, 2]))
-    >>> tag = F.tensor([1, 1, 0, 2, 0, 1, 1, 0])
+    >>> tag = torch.tensor([1, 1, 0, 2, 0, 1, 1, 0])
     >>> g_sorted = dgl.sort_csr_by_tag(g, tag, tag_type='edge')
     >>> g_sorted.adj(scipy_fmt='csr').nonzero()
     (array([0, 0, 0, 0, 0, 1, 1, 1], dtype=int32), array([2, 4, 0, 1, 3, 2, 0, 1], dtype=int32))
@@ -2995,6 +2996,7 @@ def sort_csc_by_tag(g, tag, tag_offset_name='_TAG_OFFSET', tag_type='node'):
     ``tag_type`` is ``node``.
 
     >>> import dgl
+    >>> import torch
     >>> g = dgl.graph(([0,1,2,3,4,0,1,2],[0,0,0,0,0,1,1,1]))
     >>> g.adjacency_matrix(scipy_fmt='csr', transpose=True).nonzero()
     (array([0, 0, 0, 0, 0, 1, 1, 1], dtype=int32),
@@ -3013,9 +3015,8 @@ def sort_csc_by_tag(g, tag, tag_offset_name='_TAG_OFFSET', tag_type='node'):
 
     ``tag_type`` is ``edge``.
 
-    >>> from dgl import backend as F
     >>> g = dgl.graph(([0,1,2,3,4,0,1,2],[0,0,0,0,0,1,1,1]))
-    >>> tag = F.tensor([1, 1, 0, 2, 0, 1, 1, 0])
+    >>> tag = torch.tensor([1, 1, 0, 2, 0, 1, 1, 0])
     >>> g_sorted = dgl.sort_csc_by_tag(g, tag, tag_type='edge')
     >>> g_sorted.adj(scipy_fmt='csr', transpose=True).nonzero()
     (array([0, 0, 0, 0, 0, 1, 1, 1], dtype=int32), array([2, 4, 0, 1, 3, 2, 0, 1], dtype=int32))