CRA Tax Rules Database MVP - October 2025

Production SQLite database with CRA business expense rules.

Database Statistics

• Total Chunks: 63 searchable expense rules
• Schema Version: 1.0
• Data Version: 2025.10.23
• Embedding Model: BGE-small-en-v1.5 (384 dimensions)
• Source Documents: CRA T4002 Business and Professional Income Guide
• Source Files: t4002-5.html
• Lineage Coverage: 100% (classic parser + adjudicator)
• Expense Types: 16 unique categories
• Search Indices: FTS5 full-text + sqlite-vec semantic search

File Verification

SHA256 Checksum:

960601b0e7a1b73e59ce8f857b603516382341ba22729da9d90858bbf0f6125b

Verify download:

shasum -a 256 cra_rules.db
# Should match: 960601b0e7a1b73e59ce8f857b603516382341ba22729da9d90858bbf0f6125b

Usage

Install the library:

pip install qe-tax-rag

The database will be downloaded automatically on first use:

import qe_tax_rag as qe

# Downloads database from this GitHub Release
qe.init()

# Search expense rules
results = qe.search(
    query="restaurant meals for client meetings",
    province="BC",
    expense_types=["meals"]
)

Quality Validation

This database has been validated through comprehensive testing:

• Search Success Rate: 10/10 queries (100% - exceeds 70% threshold)
• Lineage Traceability: 100% coverage with source document tracking
• Integrity Checks: All database constraints verified
• Embedding Quality: All 63 chunks have valid 384-dim vectors

See PR #44 and
PR #45 for full validation
reports.

Distribution Model

• Code: Distributed via PyPI (pip install qe-tax-rag)
• Database: Downloaded from GitHub Releases on first use
• Versioning: Schema version + data version stored in database metadata
• Integrity: SHA256 checksums verified on download

Disclaimer

⚠️ IMPORTANT: NOT FINANCIAL OR TAX ADVICE

This software is provided for informational purposes only and is not a substitute for
professional financial or tax advice. Always consult with a qualified tax professional
before making financial decisions.

Support

• Documentation: https://github.com/manonja/quickExpense-rag
• Issues: https://github.com/manonja/quickExpense-rag/issues
• Examples: See examples/ directory in repository