OCR function, PNG works except for PDF files

[fsouren]

Thanks for your reply! I'll dig into it.

But on the other hand, i'm probably not the only one trying to index Dutch text PDF files i guess. (English text PDF works great).
I just can't figure out where it goes wrong.

[jllort]

Should debug the temp files created before be executed with OCR. Two weeks ago we've released portable dev environment http://sourceforge.net/projects/openkmportabledev/ my suggestion is download it, set some breakpoint into pdf text extractor, and step by step, take a look about what's happening, specially on tmp files. ( upload only one document and from crontab -> force indexing ).

[fsouren]

Could you try one more thing for me?
I have 2 scans, doc1.pdf and doc3.pdf. doc1.pdf works, doc3.pdf doesn't.

What could be the difference?

http://www.famsouren.nl/doc1.pdf
http://www.famsouren.nl/doc3.pdf

doc3.pdf works if i first convert it manually to PNG, then upload the PNG file.

[jllort]

I've test in our online demo and seems there is going right. I attach here the text extracted.

Do you got the last OpenKM version ( the nighly build, because there're we've corrected some issues http://integration.openkm.com/ and here information about migration http://wiki.openkm.com/index.php/Migration_Guide).

[fsouren]

I've tried what you said, and upgraded to build 8186.
But still i get a lot of garbage when indexing doc3.pdf.

Sadly enough i should let it go i guess, i just can't seem the get it working

[fsouren]

Is it possible to drop the settings from the demo site here? So i can compare them.
Can't seem to view them when logging in as a demo user.

[jllort]

Demo is based on professional version, is not he community ( both versions have a similar base, but are quite different ).

For what you told us, with nightly build you got exactly the same problem no ? Can you post here a text file with extracted contents ?

[fsouren]

Yes, exact the same problem.

[fsouren]

Anyone?

[pavila]

So you've tested with a recent nightbuilt, haven't you?

[fsouren]

Yes i did. I even did a clean install with Ubuntu 14.04 en OpenKM nightly.
The only thing i did was install LibreOffice and ImageMagick, then OpenKM and replace OpenKM.war with a nightly one.

[pavila]

I've made some improvements to PDF text extraction, please try with tonight nightbuild.

Check you have installed Tesseract and configured the com.openkm.extractor.Tesseract3TextExtractor in registered.text.extractors. If present, remove com.openkm.extractor.CuneiformTextExtractor.

[fsouren]

Thanks for looking into this!

So i should wait till tomorrow? Build 8189 still ins't working for me.
Still the same text as uploaded in doc3.zip.

[pavila]

Wait until tomorrow to generate a new build.

[fsouren]

I installed the new build and it's working great now!
Could you try to explain what the underlying problem was? (in simple english please )