Add support for translating PDF files

[zwj2004]

Currently, the translation server supports .txt and .epub formats. Please add functionality to allow users to upload and translate PDF files as well. This feature would enhance usability for those who work with PDF documents and expand the application's supported formats.

Key requirements:

• Accept PDF uploads via the web interface and API
• Integrate PDF parsing and text extraction (e.g., using PyPDF2 or pdfplumber)
• Ensure extracted text is sent for translation in chunks, similar to existing formats
• Save translated output as a PDF or .txt for download
• Update documentation to reflect PDF support

This enhancement would be valuable for users with books, articles, or other documents in PDF format.

[grigio]

There are tool to convert PDF in epub and back, maybe some formatting is lost or became weird in an automatic PDF translation

[hydropix]

Thank you for raising this issue! PDF support is undoubtedly a very useful feature.

I have studied PDF processing, and it is quite complex. There is no reliable method for cleanly converting PDFs to other formats while perfectly preserving the layout. https://github.com/datalab-to/marker is probably the best solution available, using sophisticated methods that require significant computing resources, but even this fails to preserve the formatting accurately. The advantage, however, is that it can convert to HTML files. EPUB files are a series of HTML files. If you can accept that the layout will be slightly different but still readable, this is probably a good solution.