Skip to content

Latest commit

 

History

History
42 lines (30 loc) · 2.29 KB

README.md

File metadata and controls

42 lines (30 loc) · 2.29 KB

🔥MOJO REFACTOR w/ 🤗 Hugging Face "urchade/gliner_base"

Explorations by Aylex Riom (Snyata)

##### THIS IS A WORK IN PROGRESS TO BE COMPLETED - LAST UPDATED SAT 6 APR 2024

Original Knowledge Graph Creator command line tool: kgcreator

Below is the original details.

PyPI Changelog Tests License

Knowledge Graph Creator: converts text to RDF triples.

The Knowledge Graph Creator (kgcreator) is a tool for automating the generation of RDF data for Knowledge Graphs from raw text data read from an input directory path.

The Knowledge Graph Creator creates an output file containing RDF triples suitable for loading into any linked data/semantic web data store.

This Python command line utility is one of the example programs in my book Practical Python Artificial Intelligence Programming that is available on Leanpub.com or available (in the future, this book is under development) to read for free on my web site https://markwatson.com.

Notes on generating Neo4J graph data

Start by identifying:

  • Names of entities (node labels).
  • Names of relationships.
  • Names of properties for nodes and relationships.
  • Constraints to be defined.
  • Indexes required.
  • The most important queries?

Additional KGLiNER functionality

  • Contextual labelling of the entities based on the structured feedback of an LLM using the Portkey AI Gateway.

  • Optimized chunking strategies depending on the type of use case. Implementation here is ambiguous.

  • Additional Metadata is applied to the Graph as it is developed.

  • Rust based Petgraph Knowledge Graph (TBC)

  • More will come.

    Contributors welcome: [email protected]