Implicit Word Network¶

A Python package for extracting and exploring (contextual) implicit entity networks from text corpora.

Implicit entity networks represent a document collection as a cooccurrence graph of named entities and terms: entities that are mentioned close to each other are connected, and the strength of the relation decays with the distance of the mentions. The model was introduced by Spitz & Gertz and powers entity-centric corpus exploration tools such as ECCE.

Example implicit entity network

Entity layer of the network extracted from the bundled example corpus (node size = mentions, edge width = ω).

Quick Start¶

Installation¶

pip install implicit-word-network

# with spaCy NER
pip install "implicit-word-network[spacy]"
python -m spacy download en_core_web_sm

# with zero-shot GLiNER v2.5 NER
pip install "implicit-word-network[gliner]"

Basic Usage¶

from implicit_word_network import Corpus, GLiNEREntityExtractor, ImplicitNetworkPipeline

corpus = Corpus.from_txt("documents.txt")  # one document per line

extractor = GLiNEREntityExtractor(labels=["person", "organization", "location"])
pipeline = ImplicitNetworkPipeline(extractor, window=2)
network = pipeline.run(corpus, show_progress=True)

print(network.summary())
for edge in network.edges(top_k=10):
    print(edge.source.text, "--", edge.target.text, round(edge.weight, 2))

graph = network.to_networkx()  # continue with NetworkX

Documentation¶

Getting Started - Installation and first steps
Theory - The implicit entity network model
Tutorials - Building, querying, extracting, clustering, scaling
Examples - Scripts and an executed notebook
CLI Reference - Command-line interface
API Reference - Complete API documentation
Development - Contributing and development setup

Author¶

Julian Schelb - University of Konstanz

Citation¶

If you use this package in your research, please cite the underlying models:

@inproceedings{schelb2022ecce,
  title     = {ECCE: Entity-centric Corpus Exploration Using Contextual Implicit Networks},
  author    = {Schelb, Julian and Ehrmann, Maud and Romanello, Matteo and Spitz, Andreas},
  booktitle = {Companion Proceedings of the Web Conference 2022 (WWW '22 Companion)},
  year      = {2022},
  doi       = {10.1145/3487553.3524237}
}

@inproceedings{spitz2016load,
  title     = {Terms over LOAD: Leveraging Named Entities for Cross-Document Extraction and Summarization of Events},
  author    = {Spitz, Andreas and Gertz, Michael},
  booktitle = {SIGIR '16},
  year      = {2016},
  doi       = {10.1145/2911451.2911529}
}

@inproceedings{spitz2018entangled,
  title     = {Exploring Entity-centric Networks in Entangled News Streams},
  author    = {Spitz, Andreas and Gertz, Michael},
  booktitle = {Companion of the The Web Conference 2018 (WWW '18 Companion)},
  year      = {2018},
  doi       = {10.1145/3184558.3188726}
}

License¶

This project is licensed under the MIT License.