OCR for PDF always return empty string

[aleksapetkovic25]

Why google always return empty string for pdf document?

[DninoAdnane]

Hello @aleksapetkovic25 ,

Can you provide more information about the issue you are having, for example a sample file we can test with or the language in which the file is written !

Are you always having this issue, or it only happens sometimes ?

[aleksapetkovic25]

Hello @DninoAdnane , I always have this problem. Yesterday I took it to test this OCR for application at work. For image this work good, but for PDF not. I use google provider, and always when I try this on your website, postman and in my project, response was empty string. Language was Serbian(latin), but after that I tried also with English. PDF when I used file_url: https://s29.q4cdn.com/175625835/files/doc_downloads/test.pdf PDF when I used file: file in attachment, img of local pfd https://prnt.sc/Ys3WGvUCAiT3 Also I give you image of response from postman. Same thing is from your website and from my project. https://prnt.sc/xJEf-eE_KcHn Greeting, Aleksa! :) сре, 6. мар 2024. у 17:06 DninoAdnane ***@***.***> је написао/ла:

…

Hello @aleksapetkovic25 <https://github.com/aleksapetkovic25> , Can you provide more information about the issue you are having, for example a sample file we can test with or the language in which the file is written ! Are you always having this issue, or it only happens sometimes ? — Reply to this email directly, view it on GitHub <#147 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AXYOVW75RB73CJIEXZKKGNDYW45KBAVCNFSM6AAAAABEIY34TKVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTSOBRGIYTQMJTGA> . You are receiving this because you were mentioned.Message ID: ***@***.***>

[DninoAdnane]

Hello @aleksapetkovic25 ,

Really sorry for my late response, I was absent these recent days, we will investigate this concert and let you know as soon as possible.

As a first thought, most (if it's not all) ocr providers does not support PDF files, however, it's still weird that google returns an empty response instead of a bad request.

In the meantime, I can suggest you to use ocr_async which is more suitable for PDF files,

Best regards,

Adnane