fix(docsummarizer): standardize DOCX processing with single-pass conv…

…ersion

- Removed split processing for DOCX files due to issues with chunked text extraction.
- Updated conversion logic to enforce single-pass conversion for improved stability and accuracy.

fix(docsummarizer): standardize DOCX processing with single-pass conv…

…ersion

- Removed split processing for DOCX files due to issues with chunked text extraction.
- Updated conversion logic to enforce single-pass conversion for improved stability and accuracy.

feat(docsummarizer): add OCR fallback for PDF backend with text sanit…

…ization and retry support

- Introduced a fallback OCR backend (`OcrPdfBackend`) for handling garbled text in PDFs, with configurable support for OCR-based extraction.
- Enhanced text processing by sanitizing gibberish encodings in markdown across summarization workflows.
- Updated conversion logic to allow backend retry mechanisms, ensuring robust handling of corrupt text layers.

feat(docsummarizer): simplify ONNX model download and optimize embedd…

…ing budget

- Removed interactive Spectre progress bar during ONNX model downloads to prevent concurrent display issues.
- Adjusted embedding budget parameters for improved coverage and performance on small-chunk documents.
- Updated verbose logging to better track embedding progress and status.

feat(docsummarizer): enhance mode reasoning and configuration insights

- Refined summarization mode selection logic to improve user transparency while preserving auto-mode behavior.
- Added detailed reasoning for backend configuration, including embedding, summarization, PDF processing, and vector storage.
- Updated service detection output with actionable tips and clearer guidance for enabling additional features.
- Improved verbose configuration display for better debugging and understanding of system decisions.

test(config): update OnnxConfig test to reflect model directory chang…

…e to app base directory

- Changed OnnxConfig test to assert the model directory is based in the application directory instead of the user profile for better portability.
- Updated ChunkCacheService tests to clarify flaky CI behavior and recommend manual execution.

test(config): update OnnxConfig test to reflect model directory chang…

…e to app base directory

- Changed OnnxConfig test to assert the model directory is based in the application directory instead of the user profile for better portability.
- Updated ChunkCacheService tests to clarify flaky CI behavior and recommend manual execution.

docs(datasummarizer): clarify trust model, heuristics, and safety gua…

…rantees

- Added "Who this is for" section with clear use cases and non-goals
- Moved heuristics table earlier with threshold documentation and "flag, not prove" clarification
- Documented SQL safety constraints (read-only, no COPY/ATTACH/INSTALL/EXPORT, 20-row limit)
- Clarified Registry vs Store distinction with their respective purposes
- Added advisory note that auto-generated constraints require review before CI enforcement
- Highlighted .NET 10 foundation for high-performance local analytics

docs(datasummarizer): clarify trust model, heuristics, and safety gua…

…rantees

- Added "Who this is for" section with clear use cases and non-goals
- Moved heuristics table earlier with threshold documentation and "flag, not prove" clarification
- Documented SQL safety constraints (read-only, no COPY/ATTACH/INSTALL/EXPORT, 20-row limit)
- Clarified Registry vs Store distinction with their respective purposes
- Added advisory note that auto-generated constraints require review before CI enforcement
- Highlighted .NET 10 foundation for high-performance local analytics

docs(datasummarizer): clarify trust model, heuristics, and safety gua…

…rantees

- Added "Who this is for" section with clear use cases and non-goals
- Moved heuristics table earlier with threshold documentation and "flag, not prove" clarification
- Documented SQL safety constraints (read-only, no COPY/ATTACH/INSTALL/EXPORT, 20-row limit)
- Clarified Registry vs Store distinction with their respective purposes
- Added advisory note that auto-generated constraints require review before CI enforcement
- Highlighted .NET 10 foundation for high-performance local analytics