v0.0.2

feyninc/pulpiev0.0.2Jul 1, 2026by chonknick

AI Summary

This is the first release under the feyninc organization, focusing on improving developer experience (DX) for extractors and pipelines, adding Apple MPS support, and ensuring byte-parity with MinerU-HTML's simplification and reconstruction.

Key Highlights

First release under feyninc org
Improved DX: markdown cleanup, error propagation, extract_batch, and dataclass results
MinerU-HTML parity: Port of simplify/reconstruct with byte-parity testing
Apple MPS auto-detection
Model ID updates to feyninc/pulpie-orange-{small,base,large}

Breaking Changes

Model IDs now resolve to feyninc/pulpie-orange-{small,base,large} (requires code update for older versions)

New Features

Apple MPS auto-detection
extract_batch method
Dataclass results for extractors

Full Release Notes

First release under the feyninc org.

- Ships `pulpie.markdown` (the `[markdown]` extra now matches the docs).
- Extractor/Pipeline DX fixes: markdown cleanup, error propagation, `extract_batch`, dataclass results.
- MinerU-HTML-parity `simplify`/`reconstruct` port (byte-parity tested against the vendored oracle).
- Apple MPS auto-detection.
- Model IDs now resolve to `feyninc/pulpie-orange-{small,base,large}`.

Install: `pip install "pulpie[markdown]"`