v0.5.1 - CSV refresh, fund/listed precedence fix, robust EN/JP loader

Changes

• Fund-precedence fix: listed-company status now wins over fund-registry membership in EntityClassifier. Credit Saison (E03041) and JAFCO (E04806) no longer misclassified as FUND.
• EdinetcodeDlInfo and FundcodeDlInfo CSVs refreshed to 2026-04-13. processed_companies.csv regenerated. Old March 9 CSVs removed.
• Robust EN/JP loader: columns are resolved by header name via known English/Japanese aliases, so both FSA CSV variants work transparently. Schema changes fail loudly with a clear error naming the missing field, rather than silently returning data from the wrong column.
• Industry normalization: new translate_industry_to_english() helper. Entity records now expose both industry (English, stable) and industry_jp(raw source value). Company search matches either language.
• New scripts/refresh_csvs.py: downloads fresh CSVs from FSA and installs them into the package data directory with the right date suffix.

612 tests pass.

v0.5.0 — Complete EDINET document type coverage

Typed parsers for all known EDINET document types (up from 9). Amendment support, complete doc type registry,
fetch_document(type=) for PDF/HTML downloads.

v0.4.3

• fetch_and_parse(doc_id, doc_type_code) - parse documents by ID directly
• Entity.industry - industry classification now exposed on Entity objects
• Two new tests, 348 passing.

v0.4.2

Fixed

• IFRS financial extraction for ~6% of listed companies
• Context matching bug that could leak non-consolidated data into consolidated results
• IFRS Summary tier added (revenue, operating income, net income, assets, equity, EPS, ROE)
• Expanded IFRS FS fallback map (balance sheet detail, income detail, debt)

v0.4.1

What's New

• Typed report parsers for all seven EDINET doc types supported - doc.parse() now returns SecuritiesReport, LargeHoldingReport, TreasuryStockReport, etc. with typed fields
• Shared extraction framework (extraction.py) for consistent XBRL/CSV parsing
• Lighter install - llm and pydantic moved to optional deps (pip install edinet-tools[analysis])
• Doc type registry expanded - added treasury stock (220/230)
• Test suite overhaul - removed redundant/stale tests, added parser extraction tests
• Updated demo.py showcasing typed parser API

Breaking Changes

• analysis.py removed
• Test dir renamed test/ -> tests/

v0.4.0

Tender offer parser (Doc 240/250), API timeout + exponential backoff, updated entity data

v0.3.0

What's New

Treasury Stock Reports

New TreasuryStockReport parser for doc types 220/230 — extracts shares acquired/disposed, acquisition price, and holding totals.

IFRS Cash Flow Support

Securities report parser now handles IFRS-format cash flow statements in addition to Japanese GAAP.

Debt Field Extraction

Securities reports now extract short-term and long-term debt balances from balance sheet data.

Other Improvements

• JST timezone utility for consistent datetime handling

Contributors

• @jolks — balance sheet enhancements (#6)