Add PDF loading support and update dependencies

[iam-tsr]

🧠 Context

This pull request adds support for PDF document loading to the mofa-foundation RAG pipeline, enabling users to extract and process text from PDF files as part of their retrieval-augmented generation workflows. The implementation introduces a new PdfLoader (behind the pdf feature flag), integrates it into the pipeline and documentation, and provides tests and usage examples. Additionally, a small bug fix is included in the Python example.

PDF Document Loading Support

• Added a new PdfLoader struct implementing the DocumentLoader trait, allowing extraction of text from PDF files using the pdf-extract crate. The loader handles file extension checks, error handling, and metadata population.
• Introduced a new LoaderError::PdfParseError variant for robust error reporting when PDF parsing fails.
• Enabled the pdf feature in mofa-foundation and rag_pipeline examples, updating Cargo.toml files and documentation to describe PDF support and usage.

Examples and Tests

• Added integration and unit tests for PdfLoader covering extension checks and metadata, and included a demonstration of PDF loading and chunking in the rag_pipeline example.

Bug Fix

• Fixed a typo in the Python analyze.py script where "analyses.append("sentiment")" was missing a closing quote and bracket, correcting it to result["analyses"].append("sentiment").