Fix CI (stellargraph#1840)

This fixes numerous issues causing CI to fail. None of them stopped the models
working.

- Update the `MovieLens` dataset to specify an encoding
  (https://en.wikipedia.org/wiki/ISO/IEC_8859-1 in particular), because the
  file contains some invalid UTF-8, in the form of rows in `u.item` such as
  `Misérables, Les (1995)|...`, where the `é` is encoded as `E9`, instead of
  `C3 A9` (see https://en.wikipedia.org/wiki/%C3%89#Character_mappings). This
  started failing when pandas 1.2.0 was released (it passed with 1.1.5), and
  the relevant changelog
  (https://pandas.pydata.org/pandas-docs/stable/whatsnew/v1.2.0.html#i-o) might
  be (unclear, though):

   > to_csv() and read_csv() did not honor compression and encoding for
   > path-like objects that are internally converted to file-like objects
   > (GH35677, GH26124, GH32392)

- Remove the `with_labels` argument from `nx.draw_networkx_nodes` calls
  (https://networkx.org/documentation/stable/reference/generated/networkx.drawing.nx_pylab.draw_networkx_nodes.html#networkx.drawing.nx_pylab.draw_networkx_nodes),
  which always did nothing and were just being grouped into the unused `**kwds`
  dictionary in NetworkX 2.4 and earlier. NetworkX 2.5 removes this dictionary
  and so the argument cannot be specified.  (I suspect it was inherited from
  `nx.draw_networkx`
  (https://networkx.org/documentation/stable/reference/generated/networkx.drawing.nx_pylab.draw_networkx.html),
  which does support `with_labels`.)  Relevant changelog
  (https://networkx.org/documentation/stable/release/release_2.5.html#api-changes):

  > all keywords are now checked for validity in nx.draw and friends

  Fixes stellargraph#1798

- Switch to appending directories to path via the file referred to by
  `$GITHUB_PATH`, because GitHub stopped using the special printed `::add-path`
  command. Relevant changelogs:
  - https://github.blog/changelog/2020-10-01-github-actions-deprecating-set-env-and-add-path-commands/
  - https://github.blog/changelog/2020-11-09-github-actions-removing-set-env-and-add-path-commands-on-november-16/

- Force `__file__` to be relative to the root of the repo
  `scripts/demo_indexing.py`, because it changed to be the absolute path,
  rather than the path relative to the root of the repo, unsure why.

- Switch `validation_data=[data, labels]` to `validation_data=(data, labels)`
  (a tuple) in GCN-LSTM demo, because TensorFlow 2.4 changed the behaviour of
  `model.fit(..., validation_data=...)`, to require a tuple to get that
  behaviour (the list version tries to use the list as inputs). Potentially
  relevant changelog
  (https://github.com/tensorflow/tensorflow/releases/tag/v2.4.0) (unclear):

  > For Keras model, the individual call of Model.evaluate uses no cached data
  > for evaluation, while Model.fit uses cached data when validation_data arg
  > is provided for better performance.

- Remove the `port=None` (and `user` and `password`) arguments from the
  `py2neo.Graph` constructor call, because py2neo 2020.0.0 changed how the
  `Graph(...)` constructor handles them: in particular, `port=None` tries to
  use `None` as the port, rather than the default value. py2neo doesn't seem to
  publish a changelog. Fixes stellargraph#1743.

- Use `python-version: '3.8'` for `setup-python` on CI, because Python 3.9 was
  released, and the `'3.x'` constraing was using it. There isn't a TensorFlow
  release for 3.9 yet, so installation would fail in the general checkers using
  this constraint.

- Switch to explicit `runner: ubuntu-18.04` on CI, because Github upgraded the
  `ubuntu-latest` runner version from Ubuntu 18.04 to Ubuntu 20.04, which
  caused two steps to fail:

  - the copyright header checking starts failing for an unknown and
    non-described reason in the `find` invocation
  - building the documentation fails the spell-checking step, seemingly because
    the `en_AU` dictionary cannot be found, and instead it's using US spelling
    for `visualisation` etc.

  Relevant changelog:
  https://github.blog/changelog/2020-10-29-github-actions-ubuntu-latest-workflows-will-use-ubuntu-20-04/

.github/workflows/ci.yml

@@ -293,7 +293,7 @@ jobs:
         run: |
           python3 -m pip install --upgrade pip
           python3 -m pip install black==19.10b0
-          echo ::add-path::$HOME/.local/bin
+          echo "$HOME/.local/bin" >> $GITHUB_PATH
       - name: Check formatting
         run: |
           if ! black --check --diff . ; then
@@ -319,7 +319,7 @@ jobs:
           filter_mode: file
 
   copyright:
-    runs-on: ubuntu-latest
+    runs-on: ubuntu-18.04
     steps:
       - uses: actions/checkout@v2
 
@@ -369,7 +369,7 @@ jobs:
       - uses: actions/checkout@v2
       - uses: actions/setup-python@v2
         with:
-          python-version: "3.x"
+          python-version: "3.8"
 
       - name: Install requirements
         run: |
@@ -386,7 +386,7 @@ jobs:
       - uses: actions/checkout@v2
       - uses: actions/setup-python@v2
         with:
-          python-version: "3.x"
+          python-version: "3.8"
 
       - name: Install requirements
         run: |
@@ -450,7 +450,7 @@ jobs:
       - uses: actions/checkout@v2
       - uses: actions/setup-python@v2
         with:
-          python-version: "3.x"
+          python-version: "3.8"
 
       - name: Install requirements
         run: |
@@ -461,14 +461,14 @@ jobs:
         run: python scripts/demo_indexing.py --action=compare
 
   check-documentation:
-    runs-on: ubuntu-latest
+    runs-on: ubuntu-18.04
     env:
       sphinx_opts: "-W --keep-going"
     steps:
       - uses: actions/checkout@v2
       - uses: actions/setup-python@v2
         with:
-          python-version: "3.x"
+          python-version: "3.8"
 
       - name: Install pandoc
         run: |

demos/basics/loading-saving-neo4j.ipynb

@@ -116,8 +116,8 @@
     "\n",
     "default_host = os.environ.get(\"STELLARGRAPH_NEO4J_HOST\")\n",
     "\n",
-    "# Create the Neo4j Graph database object; the parameters can be edited to specify location and authentication\n",
-    "neo4j_graph = py2neo.Graph(host=default_host, port=None, user=None, password=None)"
+    "# Create the Neo4j Graph database object; port, user, password parameters can be add to specify location and authentication\n",
+    "neo4j_graph = py2neo.Graph(host=default_host)"
    ]
   },
   {

demos/community_detection/attacks_clustering_analysis.ipynb

@@ -882,9 +882,7 @@
     "\n",
     "fig, ax = plt.subplots(figsize=(12, 6))\n",
     "nx.draw_networkx_edges(com_G, pos, edge_color=\"grey\")\n",
-    "nx.draw_networkx_nodes(\n",
-    "    com_G, pos, nodelist=nodes, with_labels=True, node_size=200, alpha=0.5\n",
-    ")\n",
+    "nx.draw_networkx_nodes(com_G, pos, nodelist=nodes, node_size=200, alpha=0.5)\n",
     "nx.draw_networkx_labels(com_G, pos, font_color=\"#362626\", font_size=50)\n",
     "mplleaflet.display(fig=ax.figure)"
    ]
@@ -3042,9 +3040,7 @@
     "\n",
     "fig, ax = plt.subplots(figsize=(22, 12))\n",
     "\n",
-    "nx.draw_networkx_nodes(\n",
-    "    cluster_G, pos, nodelist=nodes, with_labels=True, node_size=200, alpha=0.5\n",
-    ")\n",
+    "nx.draw_networkx_nodes(cluster_G, pos, nodelist=nodes, node_size=200, alpha=0.5)\n",
     "nx.draw_networkx_labels(cluster_G, pos, font_color=\"red\", font_size=50)\n",
     "nx.draw_networkx_edges(cluster_G, pos, edge_color=\"grey\")\n",
     "\n",

demos/connector/neo4j/cluster-gcn-on-cora-neo4j-example.ipynb

@@ -105,9 +105,8 @@
    "source": [
     "default_host = os.environ.get(\"STELLARGRAPH_NEO4J_HOST\")\n",
     "\n",
-    "# Create the Neo4j Graph database object;\n",
-    "# the arguments can be edited to specify location and authentication\n",
-    "graph = py2neo.Graph(host=default_host, port=None, user=None, password=\"pass\")\n",
+    "# Create the Neo4j Graph database object; port, user, password parameters can be add to specify location and authentication\n",
+    "graph = py2neo.Graph(host=default_host)\n",
     "neo4j_sg = Neo4jStellarGraph(graph)"
    ]
   },

demos/connector/neo4j/directed-graphsage-on-cora-neo4j-example.ipynb

@@ -107,8 +107,8 @@
     "\n",
     "default_host = os.environ.get(\"STELLARGRAPH_NEO4J_HOST\")\n",
     "\n",
-    "# Create the Neo4j Graph database object; the arguments can be edited to specify location and authentication\n",
-    "neo4j_graphdb = py2neo.Graph(host=default_host, port=None, user=None, password=None)"
+    "# Create the Neo4j Graph database object; port, user, password parameters can be add to specify location and authentication\n",
+    "neo4j_graphdb = py2neo.Graph(host=default_host)"
    ]
   },
   {

demos/connector/neo4j/load-cora-into-neo4j.ipynb

@@ -379,8 +379,8 @@
     "\n",
     "default_host = os.environ.get(\"STELLARGRAPH_NEO4J_HOST\")\n",
     "\n",
-    "# Create the Neo4j Graph database object; the arguments can be edited to specify location and authentication\n",
-    "graph = py2neo.Graph(host=default_host, port=None, user=None, password=None)"
+    "# Create the Neo4j Graph database object; port, user, password parameters can be add to specify location and authentication\n",
+    "graph = py2neo.Graph(host=default_host)"
    ]
   },
   {

demos/connector/neo4j/undirected-graphsage-on-cora-neo4j-example.ipynb

@@ -107,8 +107,8 @@
     "\n",
     "default_host = os.environ.get(\"STELLARGRAPH_NEO4J_HOST\")\n",
     "\n",
-    "# Create the Neo4j Graph database object; the arguments can be edited to specify location and authentication\n",
-    "neo4j_graphdb = py2neo.Graph(host=default_host, port=None, user=None, password=None)"
+    "# Create the Neo4j Graph database object; port, user, password parameters can be add to specify location and authentication\n",
+    "neo4j_graphdb = py2neo.Graph(host=default_host)"
    ]
   },
   {

demos/interpretability/gat-node-link-importance.ipynb

@@ -863,7 +863,6 @@
     "    node_color=positive_colors,\n",
     "    cmap=cmap,\n",
     "    node_size=positive_node_sizes,\n",
-    "    with_labels=False,\n",
     "    vmin=np.min(colors) - 0.5,\n",
     "    vmax=np.max(colors) + 0.5,\n",
     "    node_shape=\"o\",\n",
@@ -875,7 +874,6 @@
     "    node_color=negative_colors,\n",
     "    cmap=cmap,\n",
     "    node_size=negative_node_sizes,\n",
-    "    with_labels=False,\n",
     "    vmin=np.min(colors) - 0.5,\n",
     "    vmax=np.max(colors) + 0.5,\n",
     "    node_shape=\"d\",\n",
@@ -1025,7 +1023,6 @@
     "    node_color=positive_colors,\n",
     "    cmap=cmap,\n",
     "    node_size=positive_node_sizes,\n",
-    "    with_labels=False,\n",
     "    vmin=np.min(colors) - 0.5,\n",
     "    vmax=np.max(colors) + 0.5,\n",
     "    node_shape=\"o\",\n",
@@ -1037,7 +1034,6 @@
     "    node_color=negative_colors,\n",
     "    cmap=cmap,\n",
     "    node_size=negative_node_sizes,\n",
-    "    with_labels=False,\n",
     "    vmin=np.min(colors) - 0.5,\n",
     "    vmax=np.max(colors) + 0.5,\n",
     "    node_shape=\"d\",\n",

demos/interpretability/gcn-node-link-importance.ipynb

@@ -901,7 +901,6 @@
     "    node_color=positive_colors,\n",
     "    cmap=cmap,\n",
     "    node_size=positive_node_sizes,\n",
-    "    with_labels=False,\n",
     "    vmin=np.min(colors) - 0.5,\n",
     "    vmax=np.max(colors) + 0.5,\n",
     "    node_shape=\"o\",\n",
@@ -913,7 +912,6 @@
     "    node_color=negative_colors,\n",
     "    cmap=cmap,\n",
     "    node_size=negative_node_sizes,\n",
-    "    with_labels=False,\n",
     "    vmin=np.min(colors) - 0.5,\n",
     "    vmax=np.max(colors) + 0.5,\n",
     "    node_shape=\"d\",\n",
@@ -1054,7 +1052,6 @@
     "    node_color=positive_colors,\n",
     "    cmap=cmap,\n",
     "    node_size=positive_node_sizes,\n",
-    "    with_labels=False,\n",
     "    vmin=np.min(colors) - 0.5,\n",
     "    vmax=np.max(colors) + 0.5,\n",
     "    node_shape=\"o\",\n",
@@ -1066,7 +1063,6 @@
     "    node_color=negative_colors,\n",
     "    cmap=cmap,\n",
     "    node_size=negative_node_sizes,\n",
-    "    with_labels=False,\n",
     "    vmin=np.min(colors) - 0.5,\n",
     "    vmax=np.max(colors) + 0.5,\n",
     "    node_shape=\"d\",\n",

demos/interpretability/gcn-sparse-node-link-importance.ipynb

@@ -863,7 +863,6 @@
     "    node_color=positive_colors,\n",
     "    cmap=cmap,\n",
     "    node_size=positive_node_sizes,\n",
-    "    with_labels=False,\n",
     "    vmin=np.min(colors) - 0.5,\n",
     "    vmax=np.max(colors) + 0.5,\n",
     "    node_shape=\"o\",\n",
@@ -875,7 +874,6 @@
     "    node_color=negative_colors,\n",
     "    cmap=cmap,\n",
     "    node_size=negative_node_sizes,\n",
-    "    with_labels=False,\n",
     "    vmin=np.min(colors) - 0.5,\n",
     "    vmax=np.max(colors) + 0.5,\n",
     "    node_shape=\"d\",\n",
@@ -1028,7 +1026,6 @@
     "    node_color=positive_colors,\n",
     "    cmap=cmap,\n",
     "    node_size=positive_node_sizes,\n",
-    "    with_labels=False,\n",
     "    vmin=np.min(colors) - 0.5,\n",
     "    vmax=np.max(colors) + 0.5,\n",
     "    node_shape=\"o\",\n",
@@ -1040,7 +1037,6 @@
     "    node_color=negative_colors,\n",
     "    cmap=cmap,\n",
     "    node_size=negative_node_sizes,\n",
-    "    with_labels=False,\n",
     "    vmin=np.min(colors) - 0.5,\n",
     "    vmax=np.max(colors) + 0.5,\n",
     "    node_shape=\"d\",\n",

demos/time-series/gcn-lstm-time-series.ipynb

@@ -671,7 +671,7 @@
     "    batch_size=60,\n",
     "    shuffle=True,\n",
     "    verbose=0,\n",
-    "    validation_data=[testX, testY],\n",
+    "    validation_data=(testX, testY),\n",
     ")"
    ]
   },

scripts/demo_indexing.py

@@ -33,9 +33,7 @@
 LINK_DEFAULT_TEXT = "demo"
 TRUE_TEXT = "yes"
 DOC_URL_BASE = "https://stellargraph.readthedocs.io/en/stable"
-AUTOGENERATED_PROMPT = (
-    f"autogenerated by {__file__}, edit that file instead of this location"
-)
+AUTOGENERATED_PROMPT = f"autogenerated by {os.path.relpath(__file__)}, edit that file instead of this location"
 DOCS_LINK_SEPARATOR = "\n<!-- DOCS LINKS -->\n"
 
 

stellargraph/datasets/datasets.py

@@ -421,6 +421,7 @@ def load(self):
             header=None,
             names=movie_columns,
             usecols=["movie_id"] + movie_columns[5:],
+            encoding="iso-8859-1",
         )
 
         # manage the IDs