pyduck-janitor-embeddings

PyPI version Python versions License: MIT CI

Bundled embedding models for pyduck-janitor.

This companion wheel ships sentence-transformers model weights inside the package so pyduck_janitor.embed_install() works without contacting HuggingFace. It's the offline install path for air-gapped machines, CI, and anyone who'd rather not pull weights at first use.

License

  • The companion wheel's code (this pyduck_janitor_embeddings/ package) is MIT, matching the parent pyduck-janitor.
  • The bundled model weights are redistributed under their upstream license (Apache 2.0 for all-MiniLM-L6-v2); a NOTICE file inside each model directory documents the source URL and full attribution.
  • See data/embeddings/<slug>/NOTICE inside the installed wheel for the per-model trail.

Install

# Recommended — pulls the companion wheel + sentence-transformers
pip install pyduck-janitor[embeddings]

# Or install this wheel directly
pip install pyduck-janitor-embeddings

Use

import pyduck_janitor as pj

# Copies the bundled weights to the local cache. No network.
pj.embed_install()

# Inspect
pj.embed_list_installed()

Models bundled

Model Dim Size License
sentence-transformers/all-MiniLM-L6-v2 384 ~90 MB Apache 2.0

Model weights are redistributed under their original licenses; see data/embeddings/<model>/LICENSE inside the wheel for the full text.

Want a different model?

The bundled default is fine for most tasks, but pyduck_janitor.embed_install accepts any sentence-transformers-compatible model from HuggingFace Hub:

import pyduck_janitor as pj

pj.embed_install("hf:BAAI/bge-small-en-v1.5")               # better quality
pj.embed_install("hf:intfloat/multilingual-e5-small")      # multilingual
pj.embed_install("hf:org/model@sha256:abc123...")           # pinned revision
pj.embed_install("/opt/models/my-finetuned-encoder")        # local path

For gated/private models, set HF_TOKEN in your environment. See the parent project's README for the full guide.

Rebuilding the wheel (maintainers)

The weights are not committed to git. To populate them before a build:

pip install -e ".[build]"
python scripts/fetch_models.py   # downloads into src/.../data/embeddings/
python -m build                  # wheel now carries the weights

Why a separate wheel?

pyduck-janitor stays small (~50 KB) so users who only need the icu/fts text verbs don't download 90 MB of model weights. This companion package is only pulled when you ask for embeddings — via the [embeddings] extra or an explicit install.

Project links

Key dates

PyPI data
Data sourced directly from PyPI's database.
  • Released:
Latest release

1 maintainer

PyPI data
Data sourced directly from PyPI's database.
Avatar for ezraair555 from gravatar.com ezraair555

Credits

Author: EzraAir555

License expression

MIT
View SPDX License List

Requires

Python >=3.9

Provides Extra

build dev

Tags

pyduck-janitor embedding sentence-transformers huggingface

Classifiers

Report project as malware

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

pyduck_janitor_embeddings-0.1.0.tar.gz (83.5 MB view details)

Uploaded Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

pyduck_janitor_embeddings-0.1.0-py3-none-any.whl (83.5 MB view details)

Uploaded Python 3

File details

Details for the file pyduck_janitor_embeddings-0.1.0.tar.gz.

File metadata

File hashes

Hashes for pyduck_janitor_embeddings-0.1.0.tar.gz
Algorithm Hash digest
SHA256 ae45085da5018ca9e753e58d72cdb880cbc182a9bbcc7a15551f6638fd3871d3
MD5 0a3d029581172c9d726d0de9431f5ebd
BLAKE2b-256 5e59e85cc45b401b41c52f81b2fc1429b84b33e9d39bab10d83e98eda059fba2

See more details on using hashes here.

Provenance

The following attestation bundles were made for pyduck_janitor_embeddings-0.1.0.tar.gz:

Publisher: ci.yml on ezraair555/pyduck-janitor-embeddings

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

File details

Details for the file pyduck_janitor_embeddings-0.1.0-py3-none-any.whl.

File metadata

File hashes

Hashes for pyduck_janitor_embeddings-0.1.0-py3-none-any.whl
Algorithm Hash digest
SHA256 a49ad1d21a1183093a224001dc5b429bd2801b575947b6a5579eff4b74be3cc8
MD5 432a5f4a6d12e87c1ffb9a7eb8e90271
BLAKE2b-256 9dbd32373b0aab95cdaca6fd78ffd3a3b9526af4f5c1e83e1b8471a999c31683

See more details on using hashes here.

Provenance

The following attestation bundles were made for pyduck_janitor_embeddings-0.1.0-py3-none-any.whl:

Publisher: ci.yml on ezraair555/pyduck-janitor-embeddings

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

Release history Release notifications | RSS feed

This release

0.1.0 This release

2 files

Anthropic, PBC Visionary sponsor Bloomberg Visionary sponsor Hudson River Trading Visionary sponsor Meta Visionary sponsor NVIDIA Visionary sponsor Microsoft Sustainability sponsor Depot Continuous Integration AWS Cloud computing and Security Sponsor Datadog Monitoring Fastly CDN Google Download Analytics Sentry Error logging StatusPage Status page