Metadata-Version: 2.4
Name: embeddin
Version: 0.1.0
Summary: A local UI package for turning markdown or text chunk zips into embeddings
Author: Wenxi Wang
License-Expression: MIT
Project-URL: Homepage, https://example.com/embedding
Project-URL: Repository, https://example.com/embedding
Keywords: embedding,rag,vectors,streamlit,ui,sentence-transformers
Classifier: Development Status :: 3 - Alpha
Classifier: Intended Audience :: Developers
Classifier: Operating System :: OS Independent
Classifier: Programming Language :: Python :: 3
Classifier: Programming Language :: Python :: 3.9
Classifier: Programming Language :: Python :: 3.10
Classifier: Programming Language :: Python :: 3.11
Classifier: Programming Language :: Python :: 3.12
Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
Classifier: Topic :: Software Development :: Libraries :: Python Modules
Requires-Python: >=3.9
Description-Content-Type: text/markdown
License-File: LICENSE
Requires-Dist: streamlit>=1.32
Requires-Dist: sentence-transformers>=3.0.0
Requires-Dist: numpy>=1.24
Dynamic: license-file

# embedding

A local UI package for turning a zip of `.md` or `.txt` chunks into embedding vectors.

## What it does

- launches with the `embedding` command
- lets you choose an embedding model from a dropdown or type a custom model name
- reads a zip of `.md` or `.txt` files
- creates one embedding vector per file
- exports a zip with:
  - `embedding_summary.json`
  - `embedding_manifest.csv`
  - `*_embeddings.jsonl` (optional)
  - `*_embeddings.csv` (optional)
  - `*_embeddings.npz` (optional)

## Install

```bash
pip install embedding
```

## Run

```bash
embedding
```

## Suggested input

Use a zip produced after your chunking step, such as the recursive chunk zip that contains many small `.md` chunk files.

## Suggested output use

- `jsonl` for readable records and pipelines
- `csv` for spreadsheet-style inspection
- `npz` for loading embeddings directly into NumPy / Python

## Notes

- This package creates embeddings from local text files using `sentence-transformers` models.
- It does not call an LLM by itself.
- It stores one vector per input chunk file.

## Ownership note

The package metadata and copyright notice are set to Wenxi Wang. You should still verify PyPI package-name availability, trademark questions, and any legal or patent issues yourself before publishing.
