Llama Parse accuracy issue on scanned PDF with rotated pages

[sophie-aistribute]

The OCR is not working properly when the page is rotated. It would be good if a rotation detection could be performed.

For example: Apache Tika has this feature
https://cwiki.apache.org/confluence/display/tika/tikaocr