[ENH] Consistent extension sampling of CPDAG, PDAG and DAG (#102)

* Add sampling a consistent extension of a CPDAG/PDAG * Also enable conversion from DAG to CPDAG --------- Signed-off-by: Adam Li <[email protected]>
py-why · Dec 6, 2023 · ed4cb7f · ed4cb7f
1 parent 9c6a556
commit ed4cb7f
Show file tree

Hide file tree

Showing 13 changed files with 630 additions and 7 deletions.
diff --git a/.github/workflows/main.yml b/.github/workflows/main.yml
@@ -158,6 +158,7 @@ jobs:
         if: "matrix.os == 'ubuntu'"
         shell: bash
         run: |
+          sudo apt-get update
           sudo apt-get install nvidia-cuda-toolkit nvidia-cuda-toolkit-gcc
 
       - name: Run pytest # headless via Xvfb on linux

diff --git a/README.md b/README.md
@@ -76,7 +76,7 @@ To install the package from github, clone the repository and then `cd` into the
 
 Pywhy-Graphs is always looking for new contributors to help make the package better, whether it is algorithms, documentation, examples of graph usage, and more! Contributing to Pywhy-Graphs will be rewarding because you will contribute to a much needed package for causal inference.
 
-See our [contributing guide](https://github.com/py-why/pywhy-graphs/CONTRIBUTING.md) for more details.
+See our [contributing guide](https://github.com/py-why/pywhy-graphs/blob/main/CONTRIBUTING.md) for more details.
 
 # Citing
 

diff --git a/doc/api.rst b/doc/api.rst
@@ -64,6 +64,26 @@ causal graph operations.
    is_semi_directed_path
    all_semi_directed_paths
 
+:mod:`pywhy_graphs.algorithms`: Algorithms for dealing with CPDAGs
+==================================================================
+With Markov equivalence classes of DAGs in a Markovian SCM setting, we obtain
+a potentially directed acyclic graph (PDAG), which may be completed (CPDAG).
+We may want to generate a consistent DAG extension (i.e. Markov equivalent) of a CPDAG
+then we may use some of the algorithms described here. Or perhaps one may want to 
+convert a DAG to its corresponding CPDAG.
+
+.. currentmodule:: pywhy_graphs.algorithms
+
+.. autosummary::
+   :toctree: generated/
+
+   pdag_to_dag
+   dag_to_cpdag
+   pdag_to_cpdag
+   order_edges
+   label_edges
+
+
 Conversions between other package's causal graphs
 =================================================
 Other packages, such as `causal-learn <https://github.com/cmu-phil/causal-learn>`_,

diff --git a/doc/installation.md b/doc/installation.md
@@ -1,9 +1,10 @@
 Installation
 ============
 
-**pywhy-graphs** supports Python >= 3.8.
+**pywhy-graphs** closely follows the NetworkX dependencies and thus supports Python >= 3.9.
 
-## Installing with ``pip``
+Installing with ``pip``
+-----------------------
 
 **pywhy-graphs** is available [on PyPI](https://pypi.org/project/pywhy-graphs/). Just run
 
@@ -12,7 +13,8 @@ Installation
     # or if you use poetry which is recommended
     poetry add pywhy-graphs
 
-## Installing from source
+Installing from source
+----------------------
 
 To install **pywhy-graphs** from source, first clone [the repository](https://github.com/py-why/pywhy-graphs):
 

diff --git a/doc/references.bib b/doc/references.bib
@@ -17,6 +17,16 @@ @article{bareinboim_causal_2016
   pages    = {7345--7352}
 }
 
+@article{chickering2002learning,
+  title     = {Learning equivalence classes of Bayesian-network structures},
+  author    = {Chickering, David Maxwell},
+  journal   = {The Journal of Machine Learning Research},
+  volume    = {2},
+  pages     = {445--498},
+  year      = {2002},
+  publisher = {JMLR}
+}
+
 @article{Colombo2012,
   author    = {Diego Colombo and Marloes H. Maathuis and Markus Kalisch and Thomas S. Richardson},
   title     = {{Learning high-dimensional directed acyclic graphs with latent and selection variables}},

diff --git a/doc/whats_new/v0.2.rst b/doc/whats_new/v0.2.rst
@@ -30,6 +30,7 @@ Changelog
 - |Feature| Implement and test functions to convert a PAG to MAG, by `Aryan Roy`_ (:pr:`93`)
 - |API| Remove support for Python 3.8 by `Adam Li`_ (:pr:`99`)
 - |Feature| Implement a suite of functions for finding and checking semi-directed paths on a mixed-edge graph, by `Adam Li`_ (:pr:`101`)
+- |Feature| Implement functions for converting between a DAG and PDAG and CPDAG for generating consistent extensions of a CPDAG for example. These functions are :func:`pywhy_graphs.algorithms.pdag_to_cpdag`, :func:`pywhy_graphs.algorithms.pdag_to_dag` and :func:`pywhy_graphs.algorithms.dag_to_cpdag`, by `Adam Li`_ (:pr:`102`)
 
 Code and Documentation Contributors
 -----------------------------------

diff --git a/pyproject.toml b/pyproject.toml
@@ -7,7 +7,7 @@ exclude_dirs = ["tests"]
 
 [tool.black]
 line-length = 100
-target-version = ['py38']
+target-version = ['py39']
 include = '\.pyi?$'
 extend-exclude = '''
 (
@@ -102,10 +102,10 @@ readme = "README.md"
 classifiers = [
   'Development Status :: 4 - Beta',
   'License :: OSI Approved :: MIT License',
-  'Programming Language :: Python :: 3.8',
   'Programming Language :: Python :: 3.9',
   'Programming Language :: Python :: 3.10',
-  'Programming Language :: Python :: 3.11'
+  'Programming Language :: Python :: 3.11',
+  'Programming Language :: Python :: 3.12'
 ]
 keywords = ['causality', 'graphs', 'causal-inference', 'graphical-model']
 

diff --git a/pywhy_graphs/algorithms/__init__.py b/pywhy_graphs/algorithms/__init__.py
@@ -1,3 +1,4 @@
+from .cpdag import *  # noqa: F403
 from .cyclic import *  # noqa: F403
 from .generic import *  # noqa: F403
 from .multidomain import *  # noqa: F403