waccy-edgar

WACCY extension for SEC EDGAR filing parsing and pattern learning.

Status

This package provides fixture-first EDGAR/XBRL-shaped extraction, deterministic SEC companyfacts normalization, and the entry point for ExtractorRegistry discovery. Live SEC fetching is still kept separate from fixture normalization. The v0.1.0 work is tracked in:

• #6 Implement EDGAR extraction for comparable three-statement source data
• #14 Decide and implement EDGAR pattern-learning scope for v0.1.0
• v0.1.0 milestone

Installation

uv pip install waccy-edgar

Or install with the core platform:

uv pip install "waccy[edgar]"

Current Usage

from waccy.extraction import ExtractorRegistry

registry = ExtractorRegistry()
extractor = registry.get_extractor("edgar")()
print(extractor.name)

Companyfacts JSON can be normalized into WACCY fixture records before extraction:

from waccy_edgar import EdgarCompanyFactsNormalizer, EdgarExtractor

fixture = EdgarCompanyFactsNormalizer().to_fixture(companyfacts_json)
extracted = EdgarExtractor().extract({"fixture": fixture})

The normalizer selects annual 10-K facts, preserves fiscal-year and instant-fact provenance, emits FYyyyy periods, and carries partial-extraction diagnostics in fixture metadata.

Development

This package is part of the WACCY monorepo. See the main README for development setup.