feat: external tool PDF to Markdown import

[S1933]

Summary

Add PDF to Markdown import using runtime-detected external tools (Poppler + Tesseract). Users can convert any PDF in the vault into an editable Markdown note, with optional OCR for scanned pages. The source PDF is never modified.

What changed

• Added convert_pdf_to_markdown_note Tauri command (pdf_import_cmds.rs, pdf_import_extract.rs)
• Added PdfMarkdownImportDialog component with OCR mode/language selection (pdfImport.* locale keys in all 17 languages)
• Added "Convert to note" action from file preview, note-list context menu, and command palette
• Added PostHog analytics events: pdf_markdown_import_started, pdf_markdown_import_completed, pdf_markdown_import_failed
• Added pdfMarkdownImport utility (src/utils/pdfMarkdownImport.ts) with typed request/response interfaces
• Added ADR 0137 documenting the external-tool approach
• Refactored PDF import modules to eliminate circular dependencies and reduce file size (CodeScene/Codacy compliance)

Why

Users need to turn PDFs into editable Markdown notes while keeping source PDFs in the vault. Bundling a native PDF/OCR stack would increase packaging complexity before demand is proven. Runtime detection of pdfinfo, pdftotext, pdftoppm, and tesseract keeps Tolaria lean while still offering text extraction, page-by-page OCR, and a clear upgrade path.

How to test

1. Install Poppler (brew install poppler) and optionally Tesseract (brew install tesseract).
2. Launch Tolaria with a vault that contains a PDF.
3. Right-click a PDF in the note list → "Convert to note", or click the button in the file preview toolbar, or use the command palette on an active PDF tab.
4. Select OCR mode and language in the dialog, then confirm.
5. Verify a new Markdown note is created next to the PDF with frontmatter metadata, a link back to the source PDF, and extracted page content.
6. Confirm toast message on success, and error toast when Poppler/Tesseract are missing.

Screenshots

Checklist

• Code follows the project coding standards.
• Tests added or updated.
• Documentation updated.
• Existing tests pass.

[S1933]

First of all, thank you for this project!

I use this tool in my daily professional work, and it has been useful for me.

The idea for this feature comes from my own use case. I have a lot of PDF documents that I'd like to turn into a knowledge base, and I've found that Markdown files are much easier for AI agents to use than raw PDFs, especially with MCP.

This is my first feature contribution to the project, so please don't hesitate to point me in the right direction if the PR description or the implementation should be improved.