Ubuntu 14.04 docker image error: Failed to init API, possible an invalid tessdata path: /usr/local/share/

[vatsal28]

I'm using tesseract-ocr (version 4.00 beta) with tesserocr (2.2.2) and the location of tessdata folder is: /usr/local/share/

But I'm still getting the invalid tessdata path error. I've tried the following to fix it:

1. Assigned the environment variable as TESSDATA_PREFIX ='/usr/local/share/'
2. Added path='/usr/local/share/' in PyTessBaseAPI()

The location of the tessdata folder is correct but I'm still not able to use this.
Note: I'm using docker with ubuntu 14.04 image and tesserocr version is 2.2.2 and tesseract version is 4.00 beta

How do I resolve this issue?

[simonflueckiger]

can you give PyTessBaseAPI(path='/usr/local/share/tessdata/') a try?

[vatsal28]

Tried that, but that also did not work :/

[simonflueckiger]

Ok, there might be an issue with the handling of the tessdata path if you build tesserocr with a tesseract version past commit dba13db and I'm currently working on a pull request which should fix that. What build of tesseract are you using? Did you build it yourself or did you download the binaries from somewhere?

[vatsal28]

I've used the following commands to build Tesseract.

git clone https://github.com/tesseract-ocr/tesseract.git
cd tesseract
./autogen.sh
./configure --enable-debug
LDFLAGS="-L/usr/local/lib" CFLAGS="-I/usr/local/include" make
sudo make install
sudo ldconfig

For building Tesserocr, I'm using the following commands:

git clone https://github.com/sirfz/tesserocr.git
cd tesserocr
pip install .

So since I'm not specifying any particular build version, it's taking the latest one I guess.
So is this issue happening because of the latest build version?

[simonflueckiger]

Exactly, git clone https://github.com/tesseract-ocr/tesseract.git will pull the most recent commit and this might be the issue here, not entirely sure though.

Can you try the following:

pip uninstall tesserocr
git clone https://github.com/simonflueckiger/tesserocr.git
cd tesserocr
pip install .

now set TESSDATA_PREFIX to '/usr/local/share/tessdata/' and init without the path argument set in PyTessBaseAPI()

[vatsal28]

Tried this, it's not working :/ . Any other solution?

[simonflueckiger]

let's chat [tlk.io url expired]

[vatsal28]

Can we chat now ?

[simonflueckiger]

so it ended up (most likely) being a corrupted eng.traineddata file which was fixed by redownloading it:

wget -O /usr/local/share/tessdata/eng.traineddata https://github.com/tesseract-ocr/tessdata/raw/master/eng.traineddata

@sirfz this issue can be closed now

[vatsal28]

Thanks a lot!!