🗺️ Roadmap

Here are the ongoing developments for chunknorris, or features that have been identified for future development.

Each feature is developed with chunknorris' main motivation in mind: finding fast methods that require minimal computing resources. 🌱 (Sorry dear machine learning...)

Feature	Status
[PDF] Parsing of tables delimited by lines in .pdf documents	🟢
Build documentation and usage examples	🔵
[DOCX] Parsing of .docx documents to markdown	🟡
[PDF] Title tree detection when no table of contents is present in the document	🟡
[PDF] Fix typos induced by tesseract (c.f. note 1)	🟡

🔵 = ongoing 🟡 = in focus 🟢 = recently done

Note 1 : When using OCR on pdf documents, the backend used (Tesseract) may introduce typos. Some other tools, such as EasyOCR do not suffer that problem as much. But tesseract as the advantage of not relying on pytorch and running fine on CPU. Consequently, we plan on attempting to keep tesseract as a backend, but add some typo fixers.