Release 0.9.2

[jakub-walaszczyk]

Release v0.9.2

Summary

This release adds user-facing control over the chunk-size dimension of the search space and strengthens input validation for chunking parameters. Chunking constraints are now validated via Pydantic before any I/O occurs, and the minimum allowed chunk size has been lowered from 512 to 128 to support finer-grained chunking strategies.

Changes

Added

• chunk_sizes parameter on prepare_search_space_report() for constraining the chunk-size dimension of the search space (e.g. [256, 512])

Changed

• Chunking constraint validation (chunking_methods, chunk_sizes) now happens via Pydantic before any I/O, providing clearer error messages with exact field paths
• Minimum allowed chunk size lowered from 512 to 128, enabling finer-grained chunking strategies
• Duplicate values in chunking_methods and chunk_sizes are now automatically deduplicated

Migration notes

No breaking changes. The new chunk_sizes parameter is optional and defaults to platform defaults when omitted. The lowered minimum chunk size (512 → 128) only expands the accepted range — existing configurations remain valid.

Checklist

• __version__ in ai4rag/__init__.py updated to 0.9.2
• docs/about/changelog.md updated
• All tests pass (pytest)
• Docs build successfully