Merge remote-tracking branch 'origin/develop'

sknetwork-team · Mar 29, 2023 · bb3203e · bb3203e
2 parents dffd2a5 + 7372c8b
commit bb3203e
Show file tree

Hide file tree

Showing 109 changed files with 5,170 additions and 12,749 deletions.
diff --git a/.github/workflows/wheels_build.yml b/.github/workflows/wheels_build.yml
@@ -16,15 +16,15 @@ jobs:
       - uses: actions/checkout@v2
 
       - name: Build wheels
-        uses: pypa/cibuildwheel@v2.3.1
+        uses: pypa/cibuildwheel@v2.12.0
         env:
           CIBW_PLATFORM: auto
           CIBW_ARCHS: auto64
           CIBW_SKIP: cp*-musllinux*
           CIBW_ARCHS_MACOS: x86_64 arm64
           CIBW_BUILD_VERBOSITY: 3
-          CIBW_BEFORE_BUILD: "pip install -r requirements_dev.txt --user && pip install ."
-          CIBW_BEFORE_BUILD_MACOS: "pip install -r requirements_dev.txt --user && pip install ."
+          CIBW_BEFORE_BUILD: "pip install -r requirements_dev.txt && pip install ."
+          CIBW_BEFORE_BUILD_MACOS: "pip install -r requirements_dev.txt && pip install ."
           CIBW_BUILD: cp38-* cp39-* cp310-*
 
       - uses: actions/upload-artifact@v2

diff --git a/HISTORY.rst b/HISTORY.rst
@@ -2,6 +2,19 @@
 History
 =======
 
+0.29.0 (2023-03-29)
+-------------------
+
+* Change API for clustering (predict / transform)
+* Change API for classification (seeds -> labels)
+* Change API for ranking (seeds -> weights)
+* Change API for GNN (same as classifiers)
+* Remove first order methods for link prediction
+* Add nearest neighbor methods for link prediction
+* Add KNN classifier without embedding
+* Add TF-IDF
+* Solve security issues by upgrade of wheels and ipython
+
 0.28.3 (2023-01-06)
 -------------------
 

diff --git a/README.rst b/README.rst
@@ -1,6 +1,6 @@
 .. image:: https://perso.telecom-paristech.fr/bonald/logo_sknetwork.png
     :align: right
-    :width: 100px
+    :width: 150px
     :alt: logo sknetwork
 
 
@@ -21,7 +21,7 @@
 .. image:: https://img.shields.io/pypi/pyversions/scikit-network.svg
         :target: https://pypi.python.org/pypi/scikit-network
 
-Python package for the analysis of large graphs:
+Free software library in Python for the analysis of large graphs:
 
 * Memory-efficient representation of graphs as sparse matrices in scipy_ format
 * Fast algorithms
@@ -37,7 +37,6 @@ Resources
 * GitHub: https://github.com/sknetwork-team/scikit-network
 * Documentation: https://scikit-network.readthedocs.io
 
-
 Quick Start
 -----------
 
@@ -51,10 +50,17 @@ Import scikit-network::
 
     import sknetwork
 
-See our `tutorials <https://scikit-network.readthedocs.io/en/latest/tutorials/data/index.html>`_;
-the notebooks are available `here <https://github.com/sknetwork-team/scikit-network/tree/master/docs/tutorials>`_.
 
-You can also have a look at some `use cases <https://scikit-network.readthedocs.io/en/latest/use_cases/text.html>`_.
+Documentation
+-------------
+
+The documentation is structured as follows:
+
+* `Getting started <https://scikit-network.readthedocs.io/en/latest/first_steps.html>`_: First steps to install, import and use scikit-network.
+* `Reference <https://scikit-network.readthedocs.io/en/latest/reference/data.html>`_: Description of each function and object of scikit-network.
+* `Tutorials <https://scikit-network.readthedocs.io/en/latest/tutorials/data/index.html>`_: Application of the main tools to toy examples.
+* `Use cases <https://scikit-network.readthedocs.io/en/latest/use_cases/text.html>`_: More advanced examples combining several tools on specific use cases.
+* `About <https://scikit-network.readthedocs.io/en/latest/authors.html>`_: Authors, history of the library, how to contribute, index of functions and objects.
 
 Citing
 ------

diff --git a/docs/first_steps.rst b/docs/first_steps.rst
@@ -1,13 +1,13 @@
 .. _getting_started:
 
-:mod:`scikit-network` is an open-source python package for the analysis of large graphs.
+Scikit-network is an open-source python package for the analysis of large graphs.
 
 Each graph is represented by a sparse matrix in ``scipy`` format (CSR).
 
 Installation
 ------------
 
-To install :mod:`scikit-network`, run this command in your terminal:
+To install scikit-network, run this command in your terminal:
 
 .. code-block:: console
 
@@ -16,7 +16,7 @@ To install :mod:`scikit-network`, run this command in your terminal:
 If you don't have `pip`_ installed, this `Python installation guide`_ can guide
 you through the process.
 
-Alternately, you can download the sources from the `Github repo`_ and run:
+Alternately, you can download the sources from `Github`_ and run:
 
 .. code-block:: console
 
@@ -26,12 +26,12 @@ Alternately, you can download the sources from the `Github repo`_ and run:
 
 .. _pip: https://pip.pypa.io
 .. _Python installation guide: http://docs.python-guide.org/en/latest/starting/installation/
-.. _Github repo: https://github.com/sknetwork-team/scikit-network
+.. _Github: https://github.com/sknetwork-team/scikit-network
 
 Import
 ------
 
-Import :mod:`scikit-network` in Python:
+Import scikit-network in Python:
 
 .. code-block:: python
 
@@ -41,6 +41,7 @@ Graph loading
 -------------
 
 A graph is represented by its :term:`adjacency` matrix (square matrix).
+
 When the graph is bipartite, it can be represented by its :term:`biadjacency` matrix (rectangular matrix).
 
 Check our :ref:`tutorial<DataTag>` for various ways of loading a graph
@@ -61,7 +62,9 @@ Here is an example to cluster the `Karate club graph`_ with the `Louvain algorit
     adjacency = karate_club()
     algo = Louvain()
     algo.fit(adjacency)
-    labels = algo.labels_
+
+
+Please look at this :ref:`tutorial<ClusteringTag>` for a visualization of the result.
 
 .. _Karate club graph: https://en.wikipedia.org/wiki/Zachary%27s_karate_club
 .. _Louvain algorithm: https://en.wikipedia.org/wiki/Louvain_method
diff --git a/docs/reference/classification.rst b/docs/reference/classification.rst
@@ -5,23 +5,23 @@ Classification
 
 Node classification algorithms.
 
-The attribute ``labels_``  assigns a label to each node of the graph.
-
-PageRank
---------
-.. autoclass:: sknetwork.classification.PageRankClassifier
+The attribute ``labels_``  gives the label of each node of the graph.
 
 Diffusion
 ---------
 .. autoclass:: sknetwork.classification.DiffusionClassifier
 
+Nearest neighbors
+-----------------
+.. autoclass:: sknetwork.classification.KNN
+
 Propagation
 -----------
 .. autoclass:: sknetwork.classification.Propagation
 
-Nearest neighbors
------------------
-.. autoclass:: sknetwork.classification.KNN
+PageRank
+--------
+.. autoclass:: sknetwork.classification.PageRankClassifier
 
 Metrics
 -------

diff --git a/docs/reference/clustering.rst b/docs/reference/clustering.rst
@@ -56,10 +56,6 @@ Propagation
 -----------
 .. autoclass:: sknetwork.clustering.PropagationClustering
 
-K-Means
--------
-.. autoclass:: sknetwork.clustering.KMeans
-
 Post-processing
 ---------------
 .. autofunction:: sknetwork.clustering.reindex_labels

diff --git a/docs/reference/embedding.rst b/docs/reference/embedding.rst
@@ -51,8 +51,3 @@ Spring
 ------
 
 .. autoclass:: sknetwork.embedding.Spring
-
-Metrics
--------
-.. autofunction:: sknetwork.embedding.get_cosine_similarity
-
diff --git a/docs/reference/hierarchy.rst b/docs/reference/hierarchy.rst
@@ -22,10 +22,6 @@ Louvain
 
 .. autoclass:: sknetwork.hierarchy.LouvainIteration
 
-Ward
-----
-.. autoclass:: sknetwork.hierarchy.Ward
-
 Metrics
 -------
 .. autofunction:: sknetwork.hierarchy.dasgupta_cost

diff --git a/docs/reference/linkpred.rst b/docs/reference/linkpred.rst
@@ -5,32 +5,11 @@ Link prediction
 
 Link prediction algorithms.
 
-The method ``predict`` assigns a scores to edges.
+The attribute ``links_``  gives the predicted links of each node as a sparse matrix.
 
 
-First order methods
--------------------
-.. autoclass:: sknetwork.linkpred.CommonNeighbors
+Nearest neighbors
+-----------------
 
-.. autoclass:: sknetwork.linkpred.JaccardIndex
+.. autoclass:: sknetwork.linkpred.NN
 
-.. autoclass:: sknetwork.linkpred.SaltonIndex
-
-.. autoclass:: sknetwork.linkpred.SorensenIndex
-
-.. autoclass:: sknetwork.linkpred.HubPromotedIndex
-
-.. autoclass:: sknetwork.linkpred.HubDepressedIndex
-
-.. autoclass:: sknetwork.linkpred.AdamicAdar
-
-.. autoclass:: sknetwork.linkpred.ResourceAllocation
-
-.. autoclass:: sknetwork.linkpred.PreferentialAttachment
-
-
-Post-processing
----------------
-.. autofunction:: sknetwork.linkpred.is_edge
-
-.. autofunction:: sknetwork.linkpred.whitened_sigmoid
diff --git a/docs/reference/utils.rst b/docs/reference/utils.rst
@@ -26,33 +26,17 @@ Neighborhood
 .. autofunction:: sknetwork.utils.get_neighbors
 
 
-Clustering
-----------
-
-.. autoclass:: sknetwork.utils.KMeansDense
-
-.. autoclass:: sknetwork.utils.WardDense
-
 Membership matrix
 -----------------
 
 .. autofunction:: sknetwork.utils.get_membership
 
 .. autofunction:: sknetwork.utils.from_membership
 
+TF-IDF
+------
 
-Nearest-neighbors
------------------
-
-.. autoclass:: sknetwork.utils.KNNDense
-
-.. autoclass:: sknetwork.utils.CNNDense
-
-
-Co-Neighborhood
----------------
-
-.. autofunction:: sknetwork.utils.co_neighbor_graph
+.. autofunction:: sknetwork.utils.get_tfidf
 
 
 Projection
-Original file line number
+Diff line change
@@ Expand Up / @@ -51,8 +51,3 @@ Spring @@
     ------
     .. autoclass:: sknetwork.embedding.Spring
-    Metrics
-    -------
-    .. autofunction:: sknetwork.embedding.get_cosine_similarity