graph technologies

PyTextRank
a popular Python implementation of multiple textgraph graph algorithms for phrase extraction, provided as a configurable spaCy pipeline component.
- lightweight phrase extraction (unsupervised learning, CPU-based)
- leverages available NER capabilities within a pipeline
- uses: enhance indexing and document similarity; building knowledge graphs

kglab
a simple Python abstraction layer for Graph Data Science, integrating NetworkX, RAPIDS, RDFLib, Morph-KGC, pythonPSL, and more.
- provides PyData-esque interfaces to other popular graph libraries
- includes several Jupyter-based graph data science tutorials
- uses: building knowledge graphs; integrating disjoint graph technologies

- biblio – semantic bibliography entries, generated from RDF
- glossary – semantic glossary entries, generated from RDF
- apidocs – semantic apidocs supporting the Diátaxis grammar for documentation, generated as RDF from Python modules
disparity_filter
implements a disparity filter in Python, based on NetworkX, to extract the multiscale backbone of a complex weighted network (Serrano, et al., 2009)
- analogous to centrality calculated on the edges of a graph rather than its nodes; in other words, consider this as a "dual" problem of the typical graph analysis of social networks
- uses: paring down automatically-generated graphs, such as those produced by NLP methods
recommended projects within our community:
Argilla
the open-source data curation platform for LLMs, integrating closely with Hugging Face, LangChain, LlamaIndex, and more.

KùzuDB
a highly scalable, extremely fast, and very easy-to-use embeddable graph database, implementing openCypher in Python for labeled property graphs
SpanMarker
a framework for training powerful Named Entity Recognition models using familiar encoders such as BERT, RoBERTa and ELECTRA, built atop 🤗 Transformers
dotmotif
a performant, powerful query framework for identifying subgraphs or motifs in a large graph, based on NetworkX
We can help your organization develop and deploy AI applications!
Derwen provides software development, open source integration, security hardening, and performance optimization, focusing on enterprise applications which leverage natural language and graph technologies.
Rates ⊗ Contact ⊗ Policy ⊗ Newsletter
Areas of expertise within our firm include:
- graph data science applications
- developing précis, executive briefings, expert presentations
- technical co-authoring for books, industry reports, whitepapers, etc.
- methods for using natural language (LLMs, etc.) to build knowledge graphs
- leveraging W3C semantic graphs and LPG labeled property graphs together
- integrating graph visualization frameworks and customized UI/UX
- applications of probabilistic graphs
- expert software engineering+packaging in Python and C++
- developing APIs and services based on FastAPI/Pydandtic/Redis/Thespian
- parallelization and performance analysis+optimization
- building human-in-the-loop pipelines to enrich your data
- using distributed systems and cloud computing in general
Partner firms include:
Disclaimer: all trademarks, service marks, trade names, trade dress, product names, and logos appearing above are the property of their respective owners

Derwen on Medium
Derwen publication on Medium.com: newsletter summaries, articles, conferences, announcements
