Add bb document metadata update and filters

[codeforester]

Summary

Add a small read/write v2 document metadata workflow so imported generic documents can be classified and filtered without opening SQLite manually.

Scope

• Add bb documents update DOCUMENT_ID for metadata fields already present on BB_DOCUMENT.
• Support updating document type, jurisdiction code, tax year, and status.
• Add filters to bb documents list for type, jurisdiction, tax year, and status.
• Keep SQL behind DAO/service helpers rather than inline CLI command bodies.
• Keep the slice parser-free, inference-free, and extracted-text-free.
• Update README, changelog, and AI context.

Validation rules

• jurisdiction_code is normalized to uppercase when supplied.
• tax_year must be a four-digit year when supplied.
• document_status must be one of the supported v2 statuses.
• Missing document IDs fail clearly.
• An update with no metadata options fails clearly.

Out of scope

• Parser detection.
• Observation/entity inference.
• Extracted text storage.
• JSON/CSV output changes.
• Generic document inspect/report commands.

Acceptance criteria

• bb documents update DOCUMENT_ID --type bank_statement --jurisdiction US --tax-year 2025 updates the document metadata.
• bb documents list --type bank_statement --jurisdiction US --tax-year 2025 filters the document table.
• bb documents show DOCUMENT_ID reflects updated metadata.
• Existing bb documents import, list/show, and legacy commands continue working.
• Validation passes.