Computational predictive models to assist in the identification, classification, and study of scientific software.

This model is a binary classifier that predicts whether a developer and an author are the same person. It is trained on a dataset of 3000 developer-author pairs that have been annotated as either matching or not matching.

Given a set of developers and authors, we use the model on each possible pair of developer and author to predict whether they are the same person. The model returns a list of only the found matches in MatchedDevAuthor objects, each containing the developer, author, and the confidence of the prediction.

from sci_soft_models import dev_author_em
devs = [
    dev_author_em.DeveloperDetails(
        username="evamaxfield",
        name="Eva Maxfield Brown",
    ),
    dev_author_em.DeveloperDetails(
        username="nniiicc",
    ),
]
authors = [
    "Eva Brown",
    "Nicholas Weber",
]
matches = dev_author_em.match_devs_and_authors(devs=devs, authors=authors)
print(matches)
# [
#   MatchedDevAuthor(
#       dev=DeveloperDetails(
#           username='evamaxfield',
#           name='Eva Maxfield Brown',
#           email=None,
#       ),
#       author='Eva Brown',
#       confidence=0.9851127862930298
#   )
# ]

git clone https://github.com/evamaxfield/rs-graph.git
cd rs-graph
git checkout c1d8ec89
pip install -e .
rs-graph-modeling create-developer-author-em-dataset-for-annotation