PDF support through an OCR such as Tesseract

**Is your feature request related to a problem? Please describe.**
I often need to treat pdf files, when the service is deployed within a docker environment. (for named entity recognition)
A Tesseract service is also available (OCR) within my environment. A pipeline between the two would be really convenient. (But in a more general case, it should be possible to any type of OCR)
Such as :
- When we upload a pdf file, it is directly converted through a request call to the OCR_SERVER if the server exists within the environment variables. (or more globally OCR_SERVER)

- Afterwards, the task is generated just like if it was a .txt file.

**Describe the solution you'd like**
Perform a requests.post within the "uploader.py" file in order to treat pdf files.
It will take the .pdf file as an input, convert it into .txt through a requests.post, and save it as a .txt file.

**Describe alternatives you've considered**
- Monkey patching from an external application (but the feature might be useful for other people so it does not seem to be the best choice)