added pylint to the CI

.github/PULL_REQUEST_TEMPLATE.md

@@ -2,27 +2,3 @@
 
 <!-- Add a more detailed description of the changes if needed. -->
 
-## Related Issue
-
-<!-- If your PR refers to a related issue, link it here. -->
-
-## Type of Change
-
-<!-- Mark with an `x` all the checkboxes that apply (like `[x]`) -->
-
-- [ ] 📚 Examples / docs / tutorials / dependencies update
-- [ ] 🔧 Bug fix (non-breaking change which fixes an issue)
-- [ ] 🥂 Improvement (non-breaking change which improves an existing feature)
-- [ ] 🚀 New feature (non-breaking change which adds functionality)
-- [ ] 💥 Breaking change (fix or feature that would cause existing functionality to change)
-- [ ] 🔐 Security fix
-
-## Checklist
-
-<!-- Mark with an `x` all the checkboxes that apply (like `[x]`) -->
-
-- [ ] I've read the [`CODE_OF_CONDUCT.md`](https://github.com/pedugnat/dynnode2vec/blob/master/CODE_OF_CONDUCT.md) document.
-- [ ] I've read the [`CONTRIBUTING.md`](https://github.com/pedugnat/dynnode2vec/blob/master/CONTRIBUTING.md) guide.
-- [ ] I've updated the code style using `make codestyle`.
-- [ ] I've written tests for all new methods and classes that I created.
-- [ ] I've written the docstring in Google format for all the methods and classes that I used.

.github/workflows/build.yml

@@ -29,9 +29,9 @@ jobs:
         poetry config virtualenvs.in-project true
         poetry install
 
-    - name: Run style checks
+    - name: Run linting
       run: |
-        make check-codestyle
+        make lint
 
     - name: Run tests
       run: |

Makefile

@@ -44,13 +44,18 @@ check-codestyle:
 	poetry run isort --diff --check-only --settings-path pyproject.toml ./dynnode2vec ./tests
 	poetry run black --diff --check --config pyproject.toml ./dynnode2vec ./tests
 	poetry run darglint --verbosity 2 ./dynnode2vec ./tests
+	poetry run pylint ./dynnode2vec/ ./tests
+
+.PHONY: pylint
+pylint:
+	poetry run pylint ./dynnode2vec/ ./tests
 
 .PHONY: mypy
 mypy:
 	poetry run mypy --config-file pyproject.toml ./dynnode2vec ./tests
 
 .PHONY: lint
-lint: test check-codestyle mypy
+lint: check-codestyle mypy
 
 .PHONY: update-dev-deps
 update-dev-deps:

dynnode2vec/__init__.py

@@ -1,4 +1,6 @@
-"""dynnode2vec is a package to embed dynamic graphs"""
+"""
+dynnode2vec is a package to embed dynamic graphs.
+"""
 
 import sys
 from importlib import metadata as importlib_metadata
@@ -9,6 +11,7 @@
 
 
 def get_version() -> str:
+    # pylint: disable=missing-function-docstring
     try:
         return importlib_metadata.version(__name__)
     except importlib_metadata.PackageNotFoundError:  # pragma: no cover

dynnode2vec/biased_random_walk.py

@@ -1,3 +1,7 @@
+"""
+Define a BiasedRandomWalk class to perform biased random walks over graphs.
+"""
+# pylint: disable=invalid-name
 from typing import Any, Dict, Iterable, List, Union
 
 import bisect
@@ -50,7 +54,7 @@ def weighted_choice(rn: random.Random, weights: Any) -> int:
 
         This method is fastest than built-in numpy functions like `numpy.random.choice`
         or `numpy.random.multinomial`.
-        See https://stackoverflow.com/questions/24140114/fast-way-to-obtain-a-random-index-from-an-array-of-weights-in-python
+        See https://stackoverflow.com/questions/24140114/fast-way-to-obtain-a-random-index-from-an-array-of-weights-in-python  # pylint: disable=line-too-long
 
         Example: for array [1, 4, 4], index 0 will be chosen with probabilty 1/9,
         index 1 and index 2 will be chosen with probability 4/9.
@@ -69,6 +73,7 @@ def _generate_walk(
         weighted: bool,
         rn: random.Random,
     ) -> List[int]:
+        # pylint: disable=too-many-arguments, too-many-locals
         """
         Generate a number of random walks starting from a given node.
         """
@@ -94,8 +99,7 @@ def _generate_walk(
                 weights = np.ones(neighbours.shape)
 
             if (ip != 1.0) or (iq != 1.0):
-                # we update the weights according to return (p) and in-out (q)
-                # parameters
+                # we update the weights according to return (p) and in-out (q) parameters
                 mask = neighbours == previous_node
                 weights[mask] *= ip
                 mask |= np.isin(neighbours, previous_node_neighbours)
@@ -120,6 +124,7 @@ def _generate_walk_simple(
         weighted: bool,
         rn: random.Random,
     ) -> List[int]:
+        # pylint: disable=too-many-arguments
         """
         Fast implementation for the scenario where:
             - the graph is unweighted
@@ -128,7 +133,7 @@ def _generate_walk_simple(
         """
         assert ip == 1.0
         assert iq == 1.0
-        assert weighted == False
+        assert not weighted
 
         walk = [node]
 

dynnode2vec/dynnode2vec.py

@@ -1,5 +1,7 @@
+"""
+Define a DynNode2Vec class to run dynnode2vec algorithm over dynamic graphs.
+"""
 # pylint: disable=invalid-name
-
 from typing import Any, Iterable, List, Optional, Set, Tuple
 
 from collections import namedtuple
@@ -24,6 +26,8 @@ class DynNode2Vec:
     Source paper: http://www.cs.yorku.ca/~aan/research/paper/dynnode2vec.pdf
     """
 
+    # pylint: disable=too-many-instance-attributes
+
     def __init__(
         self,
         *,
@@ -37,7 +41,6 @@ def __init__(
         parallel_processes: int = 4,
         plain_node2vec: bool = False,
     ):
-        # pylint: disable=too-many-instance-attributes
         """Instantiate a DynNode2Vec object.
 
         :param p: Return hyper parameter (default: 1.0)
@@ -83,7 +86,7 @@ def __init__(
         self.parallel_processes = parallel_processes
         self.plain_node2vec = plain_node2vec
 
-        # see https://stackoverflow.com/questions/53417258/what-is-workers-parameter-in-word2vec-in-nlp
+        # see https://stackoverflow.com/questions/53417258/what-is-workers-parameter-in-word2vec-in-nlp  # pylint: disable=line-too-long
         self.gensim_workers = max(self.parallel_processes - 1, 12)
 
     def _initialize_embeddings(
@@ -160,11 +163,11 @@ def generate_updated_walks(
             # that changed compared to the previous time step
             delta_nodes = self.get_delta_nodes(current_graph, previous_graph)
 
-        BRW = BiasedRandomWalk(current_graph)
-        delta_nodes = BRW.convert_true_ids_to_int_ids(delta_nodes)
+        brw = BiasedRandomWalk(current_graph)
+        delta_nodes = brw.convert_true_ids_to_int_ids(delta_nodes)
 
         # run walks for updated nodes only
-        updated_walks = BRW.run(
+        updated_walks = brw.run(
             nodes=delta_nodes,
             walk_length=self.walk_length,
             n_walks=self.n_walks_per_node,

dynnode2vec/utils.py

@@ -1,3 +1,6 @@
+"""
+Utility file to define miscellaneous functions.
+"""
 from typing import List
 
 import random
@@ -15,6 +18,9 @@ def sample_nodes(graph: nx.Graph, k: int) -> List[int]:
 def generate_dynamic_graphs(
     n_base_nodes: int = 100, n_steps: int = 10, base_density: float = 0.01
 ) -> List[nx.Graph]:
+    """
+    Generates a list of dynamic graphs, i.e. that depend on the previous graph.
+    """
     # Create a random graph
     graph = nx.fast_gnp_random_graph(n=n_base_nodes, p=base_density)
 
@@ -25,8 +31,8 @@ def generate_dynamic_graphs(
     change_size = 1 + n_base_nodes // 10
     for _ in range(n_steps - 1):
         # remove some nodes
-        for n in sample_nodes(graph, k=change_size):
-            graph.remove_node(n)
+        for node in sample_nodes(graph, k=change_size):
+            graph.remove_node(node)
 
         # add some more nodes
         node_idx = max(graph.nodes) + 1

tests/test_biased_random_walk.py

@@ -1,3 +1,7 @@
+"""
+Test the BiasedRandomWalk class
+"""
+# pylint: disable=missing-function-docstring
 import random
 
 import numpy as np
@@ -6,27 +10,27 @@
 import dynnode2vec
 
 
-@pytest.fixture
-def graphs():
+@pytest.fixture(name="graphs")
+def fixture_graphs():
     return dynnode2vec.utils.generate_dynamic_graphs(
         n_base_nodes=30, n_steps=5, base_density=0.02
     )
 
 
 def test_init(graphs):
-    BRW = dynnode2vec.biased_random_walk.BiasedRandomWalk(graphs[0])
+    brw = dynnode2vec.biased_random_walk.BiasedRandomWalk(graphs[0])
 
     # make sure nodes ids were converted to integers
-    assert list(BRW.graph.nodes()) == list(range(BRW.graph.number_of_nodes()))
+    assert list(brw.graph.nodes()) == list(range(brw.graph.number_of_nodes()))
 
 
 def test_weighted_choice(graphs):
-    BRW = dynnode2vec.biased_random_walk.BiasedRandomWalk(graphs[0])
-    rn = random.Random(0)
+    brw = dynnode2vec.biased_random_walk.BiasedRandomWalk(graphs[0])
+    rng = random.Random(0)
     eps = 0.02
     n_try = 1_000
 
-    choices = [BRW.weighted_choice(rn, np.array([1, 4, 4])) for _ in range(n_try)]
+    choices = [brw.weighted_choice(rng, np.array([1, 4, 4])) for _ in range(n_try)]
 
     assert all(choice in [0, 1, 2] for choice in choices)
 
@@ -39,40 +43,43 @@ def test_weighted_choice(graphs):
 @pytest.mark.parametrize("iq", [1.0, 2.0])
 @pytest.mark.parametrize("weighted", [True, False])
 def test_generate_walk(graphs, ip, iq, weighted):
+    # pylint: disable=invalid-name
     # make sure that tested node has at least one neighbor
-    G = graphs[0]
-    G.add_edge(0, 1, weight=0.5)
+    graph = graphs[0]
+    graph.add_edge(0, 1, weight=0.5)
 
     # add random weights to the graph for the weighted case
     if weighted:
-        for _, _, w in G.edges(data=True):
-            w["weight"] = random.random()
+        for _, _, data in graph.edges(data=True):
+            data["weight"] = random.random()
 
-    BRW = dynnode2vec.biased_random_walk.BiasedRandomWalk(G)
-    rn = random.Random(0)
+    brw = dynnode2vec.biased_random_walk.BiasedRandomWalk(graph)
+    rng = random.Random(0)
 
-    walk = BRW._generate_walk(
-        node=0, walk_length=10, ip=ip, iq=iq, weighted=weighted, rn=rn
+    # pylint: disable=protected-access
+    walk = brw._generate_walk(
+        node=0, walk_length=10, ip=ip, iq=iq, weighted=weighted, rn=rng
     )
 
     assert isinstance(walk, list)
-    assert all(n in BRW.graph.nodes() for n in walk)
+    assert all(n in brw.graph.nodes() for n in walk)
 
 
 @pytest.mark.parametrize("p", [0.5, 1.0])
 @pytest.mark.parametrize("q", [1.0, 2.0])
 @pytest.mark.parametrize("weighted", [True, False])
 def test_run(graphs, p, q, weighted):
-    G = graphs[0]
+    # pylint: disable=invalid-name
+    graph = graphs[0]
 
     # add random weights to the graph for the weighted case
     if weighted:
-        for *_, w in G.edges(data=True):
-            w["weight"] = random.random()
+        for *_, data in graph.edges(data=True):
+            data["weight"] = random.random()
 
-    BRW = dynnode2vec.biased_random_walk.BiasedRandomWalk(G)
+    brw = dynnode2vec.biased_random_walk.BiasedRandomWalk(graph)
 
-    random_walks = BRW.run(G.nodes(), p=p, q=q, weighted=weighted)
+    random_walks = brw.run(graph.nodes(), p=p, q=q, weighted=weighted)
 
     assert all(isinstance(walk, list) for walk in random_walks)
-    assert all(n in BRW.graph.nodes() for walk in random_walks for n in walk)
+    assert all(n in brw.graph.nodes() for walk in random_walks for n in walk)