Add MATLAB compatibility and bulk document fetching

[stevevanhooser]

Summary

This PR enhances the SQLite database implementation with MATLAB compatibility for JSON serialization and adds efficient bulk document fetching capabilities.

Key Changes

• MATLAB JSON Compatibility: Added _matlab_compatible_props() method that unwraps single-element lists to scalars before JSON serialization, matching MATLAB's jsonencode behavior so DID-matlab can read Python-created databases.

• Property Normalization: Added _normalize_loaded_props() method that converts scalar values back to lists when loading documents, ensuring internal consistency regardless of the source database.

• Bulk Document Fetching: Implemented get_docs() method that fetches multiple documents in a single SQL query instead of one-at-a-time, improving performance for batch operations. Supports:

  • Single or multiple document IDs
  • Branch filtering
  • Error handling with OnMissing parameter ("error", "warn", "ignore")
  • Preserves original document ID order

• Branch-based Retrieval: Added get_docs_by_branch() convenience method to fetch all documents on a specific branch.

• Integration: Updated _do_add_doc() to apply MATLAB-compatible formatting when storing documents, and _do_get_doc() to normalize loaded properties.

Implementation Details

• The MATLAB compatibility layer specifically handles document_class.superclasses and depends_on fields, which are the fields that MATLAB converts from cell arrays to scalars.
• The bulk fetch uses parameterized SQL queries with WHERE doc_id IN (...) for safety and efficiency.
• Single document ID requests to get_docs() return a single Document object (not a list) for API convenience.

https://claude.ai/code/session_01GecDL1BqaAz3wJ7XiiAfg8