This package allows calculating page-rank and personalized page-rank via power iteration with PyTorch, which also supports calculation on GPU (or other accelerators).
As a simple example, consider this simple graph with five nodes.
Its edge list is given as
>>> import torch
>>> edge_index = torch.as_tensor(data=[(0, 1), (1, 2), (1, 3), (2, 4)]).t()We can use
>>> from torch_ppr import page_rank
>>> page_rank(edge_index=edge_index)
tensor([0.1269, 0.3694, 0.2486, 0.1269, 0.1281])to calculate the page rank, i.e., a measure of global importance.
We notice that the central node receives the largest importance score,
while all other nodes have lower importance. Moreover, the two
indistinguishable nodes 0 and 3 receive the same page rank.
We can also calculate personalized page rank which measures importance
from the perspective of a single node.
For instance, for node 2, we have
>>> from torch_ppr import personalized_page_rank
>>> personalized_page_rank(edge_index=edge_index, indices=[2])
tensor([[0.1103, 0.3484, 0.2922, 0.1103, 0.1388]])Thus, the most important node is the central node 1, nodes 0 and 3 receive
the same importance value which is below the value of the direct neighbor 4.
By the virtue of using PyTorch, the code seamlessly works on GPUs, too, and
supports auto-grad differentiation. Moreover, the calculation of personalized
page rank supports automatic batch size optimization via
torch_max_mem.
The most recent release can be installed from PyPI with:
$ pip install torch_pprThe most recent code and data can be installed directly from GitHub with:
$ pip install git+https://github.com/mberr/torch-ppr.gitContributions, whether filing an issue, making a pull request, or forking, are appreciated. See CONTRIBUTING.md for more information on getting involved.
The code in this package is licensed under the MIT License.
This package was created with @audreyfeldroy's cookiecutter package using @cthoyt's cookiecutter-snekpack template.
See developer instructions
The final section of the README is for if you want to get involved by making a code contribution.
To install in development mode, use the following:
$ git clone git+https://github.com/mberr/torch-ppr.git
$ cd torch-ppr
$ pip install -e .After cloning the repository and installing tox with pip install tox, the unit tests in the tests/ folder can be
run reproducibly with:
$ toxAdditionally, these tests are automatically re-run with each commit in a GitHub Action.
The documentation can be built locally using the following:
$ git clone git+https://github.com/mberr/torch-ppr.git
$ cd torch-ppr
$ tox -e docs
$ open docs/build/html/index.htmlThe documentation automatically installs the package as well as the docs
extra specified in the setup.cfg. sphinx plugins
like texext can be added there. Additionally, they need to be added to the
extensions list in docs/source/conf.py.
After installing the package in development mode and installing
tox with pip install tox, the commands for making a new release are contained within the finish environment
in tox.ini. Run the following from the shell:
$ tox -e finishThis script does the following:
- Uses Bump2Version to switch the version number in the
setup.cfg,src/torch_ppr/version.py, anddocs/source/conf.pyto not have the-devsuffix - Packages the code in both a tar archive and a wheel using
build - Uploads to PyPI using
twine. Be sure to have a.pypircfile configured to avoid the need for manual input at this step - Push to GitHub. You'll need to make a release going with the commit where the version was bumped.
- Bump the version to the next patch. If you made big changes and want to bump the version by minor, you can
use
tox -e bumpversion minorafter.