PyPI page
Home page
Author:
Dennis Irorere
License:
Apache-2.0
Summary:
An open-source engine for deciding when messy records of people, places, organisations, and products refer to the same real-world entity, with evidence, not just similarity scores or identifiers: decide which records describe the same person, place, organisation or product, keep the evidence, abstain when the evidence is thin, and sign the decision so a stranger can re-check it. Reads documents as well as records. Runs offline on CPU; composes with Splink, Presidio and GLiNER.
Latest version:
0.9.0
Required dependencies:
cryptography
|
h3
|
jellyfish
|
networkx
|
phonenumbers
|
pydantic
|
pyyaml
|
rapidfuzz
Optional dependencies:
anthropic
|
docling
|
duckdb
|
fastapi
|
gliner2
|
litellm
|
mcp
|
mypy
|
openai
|
presidio-analyzer
|
presidio-anonymizer
|
pyarrow
|
pymupdf
|
pypdf
|
pytest
|
pytest-asyncio
|
python-docx
|
python-multipart
|
rapidocr-onnxruntime
|
ruff
|
shapely
|
splink
|
uvicorn
Downloads last day:
0
Downloads last week:
49
Downloads last month:
345
7-day smoothing