OCRopus for Mac OSX, and a very simple GUI

[Michael Moore]

Hello,

I built OCRopus for Mac OSX 10.5.6, and wanted to share the results,
since some people seem to have had problems getting it to build. I
also compiled ImageMagick and Ghostscript so that it'll handle more
image formats.

The final result is TakOCR.app, a Mac 'dropplet' application that you
can drop images and PDFs onto and they'll be OCRed, and the hOCR
output displayed. My dad needed an OCR solution and isn't really a
command line kind of guy.

Here's the webpage:
http://stuporglue.org/tako/

And the download directory:
http://stuporglue.org/tako/downloads/v1/
* OCRopus_Full.tgz -- Tesseract, OCRopus, Imagemagick, Ghostscript --
unpack it to / and it'll expand into /usr/local/
* TakOCR.pkg -- Installer package for Mac OSX. Installs binaries and
TakOCR.app.
* TakOCR_uninstaller.command -- Bash script that uninstalls
everything TakOCR.pkg installs
* build_ocr.sh -- Script to download and build IM, GS, Tesseract,
OCRopus and needed libs on OSX. Makes needed changes to Makefile
before compiling.
* takocr.rb -- Ruby script inside TakOCR.app which sets environment
variables, splits PDFs and Tiffs if needed, and then runs the images
through OCRopus.

Feedback is welcome, thanks,
--
Michael Moore
-------------------------
Share your families' genealogy and family history books. It's easy and
free : http://bookscanned.com

[Christian Mahnke]

Hello,
> I built OCRopus for Mac OSX 10.5.6, and wanted to share the results,
> since some people seem to have had problems getting it to build. I
> also compiled ImageMagick and Ghostscript so that it'll handle more
> image formats.
>

Thanks for the great work.

> The final result is TakOCR.app, a Mac 'dropplet' application that you
> can drop images and PDFs onto and they'll be OCRed, and the hOCR
> output displayed. My dad needed an OCR solution and isn't really a
> command line kind of guy.
>
> Here's the webpage:
> http://stuporglue.org/tako/
>
> And the download directory:
> http://stuporglue.org/tako/downloads/v1/
> * OCRopus_Full.tgz -- Tesseract, OCRopus, Imagemagick, Ghostscript --
> unpack it to / and it'll expand into /usr/local/
> * TakOCR.pkg -- Installer package for Mac OSX. Installs binaries and
> TakOCR.app.
>

The Installer doesn't let me chose the desired location.

> * TakOCR_uninstaller.command -- Bash script that uninstalls
> everything TakOCR.pkg installs
> * build_ocr.sh -- Script to download and build IM, GS, Tesseract,
> OCRopus and needed libs on OSX. Makes needed changes to Makefile
> before compiling.
>

There is an small error in this file: In line 105 the script tries to
untar ocropus-0.3.1.tar.gz, which isn't downloaded before (and not even
needed).

> * takocr.rb -- Ruby script inside TakOCR.app which sets environment
> variables, splits PDFs and Tiffs if needed, and then runs the images
> through OCRopus.
>
> Feedback is welcome, thanks,
>

Maybe Thomas or Christian can add a link on ocropus.org?

Cheers,
Christian

[Thomas Breuel]

Thanks; that's great news.

We're currently trying to change the way we're doing version control to make it easier to share such contributions.  Hopefully, I can announce something in a couple of weeks.

Tom

[Michael Moore]

>> * TakOCR.pkg -- Installer package for Mac OSX. Installs binaries and
>> TakOCR.app.
>>
>
> The Installer doesn't let me chose the desired location.

I'm not a pro at Mac compilation, I'm not sure how to compile it so
that it can be relocatable, as a result, the binaries have to go in
/usr/local since that's where I compiled everything to. The TakOCR.app
itself can be moved, but the binaries can't.

I'm sure it's doable, Gimp.app and Inkscape.app do it. If someone can
point me at a tutorial on how to do it, I'd be happy to give it a
shot.

>> * build_ocr.sh -- Script to download and build IM, GS, Tesseract,
>> OCRopus and needed libs on OSX. Makes needed changes to Makefile
>> before compiling.
>>
>
> There is an small error in this file: In line 105 the script tries to
> untar ocropus-0.3.1.tar.gz, which isn't downloaded before (and not even
> needed).

Oops. I thought I'd cleaned all that out. I couldn't get 0.3.1 to
build correctly, so I switched to SVN...guess I missed a line. It's
fixed now.

> Maybe Thomas or Christian can add a link on ocropus.org?

[mwie...@online.de]

Hey! That sound great - But its not working for me :(

If I drag an Image on it (after the installation). It opens a Window -
as expected - but this windows says:

Usage: ocroscript recognize [options] input.png ... >output.hocr
Options: --tesslanguage=... Set recognition language for Tesseract --
bpnet= Use a given neural net instead of Tesseract --bpnetpath= A path
to search for the neural net --output-mode=(hocr|text) Set output
format (default: hOCR) --charboxes Output bounding boxes for
characters (Tesseract only) --langmod=... Use the given FST as
language model

noting more - If a use an PNG-Image: It opens a blank window.

:(

Any Idea?

Regards & Thanks,
Matthias

[Michael Moore]

Sorry it didn't work.

I'll look at it tonight when I get home to my Mac.

Thanks,
Michael

[Brian Chapados]

> We're currently trying to change the way we're doing version control to make
> it easier to share such contributions.  Hopefully, I can announce something
> in a couple of weeks.

I'm hoping this means that ocropus is moving to github along with the
0.4 update.

Brian

> Tom

[philip]

Doesn't work for me.

[francois...@sbktelecom.com]

I know this is a long-dead thread, but for posteriority, I got it to work by adding the following symlink:

sudo ln -s /usr/local/share/tessdata /usr/localtessdata