Insufficient error message when output file cannot be created

[Shreeshrii]

@stweil

With the new script subdirectory for the scripts traineddata,

tesseract gives error if language is given as script/Devanagari.

It works when /script is added to the tessdata-dir path and language is given as Devanagari.

I am copying my generic bash script for ruuning tesseract below, followed by console output.

[Shreeshrii]

#!/bin/bash
SOURCE="."
set -- "$SOURCE"/hin-eng.png
for img_file; do
echo -e  "\r\n File: $img_file"
for tessdir in tessdata_fast 
do    
    for lang in "Devanagari"
    do
        for psm_counter in 3 6
        do
            echo -e  "\r\n ********** $lang --oem 1 psm " $psm_counter " \r\n"
            time tesseract --tessdata-dir /mnt/c/Users/User/shree/$tessdir/script   "${img_file}" "${img_file%.*}$tessdir-$psm_counter-${lang}"  --oem 1 --psm $psm_counter -l ${lang}
        done
    done
    for lang in "script/Devanagari"
    do
        for psm_counter in 3 6
        do
            echo -e  "\r\n ********** $lang --oem 1 psm " $psm_counter " \r\n"
            time tesseract --tessdata-dir /mnt/c/Users/User/shree/$tessdir   "${img_file}" "${img_file%.*}$tessdir-$psm_counter-${lang}"  --oem 1 --psm $psm_counter -l ${lang}
        done
    done
done
done

[Shreeshrii]

#bash ./tess1.sh

File: ./hin-eng.png

********** Devanagari --oem 1 psm 3

Tesseract Open Source OCR Engine v4.0.0-beta.1-42-g3fa25 with Leptonica

real 0m7.072s
user 0m6.359s
sys 0m0.516s

********** Devanagari --oem 1 psm 6

Tesseract Open Source OCR Engine v4.0.0-beta.1-42-g3fa25 with Leptonica

real 0m6.827s
user 0m6.125s
sys 0m0.500s

********** script/Devanagari --oem 1 psm 3

Tesseract Open Source OCR Engine v4.0.0-beta.1-42-g3fa25 with Leptonica
Error during processing.

real 0m1.424s
user 0m0.781s
sys 0m0.453s

********** script/Devanagari --oem 1 psm 6

Tesseract Open Source OCR Engine v4.0.0-beta.1-42-g3fa25 with Leptonica
Error during processing.

real 0m1.411s
user 0m0.734s
sys 0m0.438s

[stweil]

That looks strange, as it works for me:

tesseract --tessdata-dir /tessdata_fast testing/devatest.png - -l script/Devanagari
CDAC-GISTYogesh A र
सभी मनुष्यों को गौरव और अधिकारों के मामले में जन्मजात स्वतन्त्रता और समानता प्रप्त हे ।

उन्हें बुद्धि और अन्तरात्मा की देन प्राप्त है और परस्पर उन्हें भाईचारे के भाव से बर्ताव करना चाहिए ।

Gargi
सभी मनुष्यों को गौरव और अधिकारों के मामले में जन्मजात स्वतन्त्रता और समानता प्राप्त हे।

उन्हें बुद्धि और अन्तरात्मा की देन प्राप्त हैऔर परस्पर उन्हें भाईचारे के भाव से बर्ताव करना चाहिए ।
JanaHindi

सभी मनुष्यों को गौरव और अधिकारों के मामले में जन्मजात स्वतन्त्रता और समानता प्रास हे ।
उन्हें बुद्धि और अन्तरात्मा की देन प्रास है और पररूपर उन्हें भाईचारे के भाव से बर्ताव करना चाहिए ।

Kalimati
सभी मनुष्यों को गोरव ओर अधिकारों के मामले में जन्मजात स्वतन्त्रता ओर समानता प्राप्त हे ।
उन्हें बुद्धि ओर अन्तरात्मा की देन प्राप्त है ओर परस्पर उन्हें भाईचारे के भाव से बर्ताव करना चाहिए ।

[Shreeshrii]

Could it be related to how I am passing the variable in bash?

[stweil]

Error during processing. is a typical case where the user interface needs to be improved, because it does not tell what the real problem was.

Could you please create a very simple test environment with a newly created traineddata directory containing nothing but traineddata/script/Devanagari.traineddata from tessdata_fast? Then run tesseract --tessdata-dir traineddata testing/devatest.png - -l script/Devanagari. That works for me.

[stweil]

Could it be related to how I am passing the variable in bash?

I see no obvious problem with the script. Try running your script with bash -x ./tess1.sh to see what is really called.

[Shreeshrii]

Thank you.

Using bash -x helped figure out the problem.

I was using $lang as part of output filename for comparing output using diff languages.

Script/Devanagari was probably looking to use a subdir because of the /.

I removed $lang and it worked fine.

Thanks. Closing the issue.

[stweil]

Yes, that's the real problem. tesseract --tessdata-dir traineddata/ testing/devatest.png abc/x -l script/Devanagari fails for me, too, with "Error during processing".

Please reopen the issue and change the title to something like "Insufficient error message when output file cannot be created".

[godofcheerup]

Hi @Shreeshrii Could I get the bash script you mentioned ?

[Shreeshrii]

@stweil Is this something you plan to address for 4.0.0?

[stweil]

It's addressed by pull request #2002.

[amitdo]

We should release 4.0.0 before #2019 comes ...:-).

[Shreeshrii]

Thanks.