text2image not found after running tesstrain.sh command

[Samruddhi Dhake]

Hello,

I am creating my own trainneddata using tesseract v4.1.1 on Windows 10.

I am referring documentation https://tesseract-ocr.github.io/tessdoc/tess4/TrainingTesseract-4.00.html

I have successfully created .box file and .lstmf file using lstmbox and lstm.train respectively.

So next step, I installed Cygwin to run tesstrain.sh command to create training data.

But I am getting below error.

$ ./src/training/tesstrain.sh --fonts_dir C:/Windows/Fonts --lang eng --linedata_only --noextract_font_properties --langdata_dir ./langdata --tessdata_dir ./tessdata --output_dir ./trainneddata

=== Starting training for language 'eng'

which: no text2image in (/usr/local/bin:/usr/bin:/cygdrive/c/Program Files/Microsoft MPI/Bin:/cygdrive/c/buildtools:/cygdrive/c/Program Files (x86)/NVIDIA Corporation/PhysX/Common:/cygdrive/c/Program Files (x86)/Intel/Intel(R) Management Engine Components/iCLS:/cygdrive/c/Program Files/Intel/Intel(R) Management Engine Components/iCLS:/cygdrive/c/Python25:/cygdrive/c/ProgramData/Oracle/Java/javapath:/cygdrive/c/Perl/site/bin:/cygdrive/c/Perl/bin:/cygdrive/c/Oracle12C_64bCli/client_1/bin:/cygdrive/c/Oracle12C_32bCli/client_1/bin:/cygdrive/c/windows/system32:/cygdrive/c/windows:/cygdrive/c/windows/System32/Wbem:/cygdrive/c/windows/System32/WindowsPowerShell/v1.0:/cygdrive/c/windows/System32/OpenSSH:/cygdrive/c/Program Files (x86)/Microsoft SQL Server/100/Tools/Binn:/cygdrive/c/Program Files/Microsoft SQL Server/100/Tools/Binn:/cygdrive/c/Program Files/Microsoft SQL Server/100/DTS/Binn:/cygdrive/c/Program Files/Microsoft/Web Platform Installer:/cygdrive/c/Program Files (x86)/Microsoft ASP.NET/ASP.NET Web Pages/v1.0:/cygdrive/c/Program Files/Microsoft SQL Server/110/Tools/Binn:/cygdrive/c/windows/system32/config/systemprofile/.dnx/bin:/cygdrive/c/Program Files/Microsoft DNX/Dnvm:/cygdrive/c/Program Files (x86)/Windows Kits/8.1/Windows Performance Toolkit:/cygdrive/c/Program Files/Microsoft SQL Server/130/Tools/Binn:/cygdrive/c/Program Files (x86)/Windows Kits/10/Windows Performance Toolkit:/cygdrive/c/Program Files (x86)/Oracle/Berkeley DB 12cR1 6.0.20/bin:/cygdrive/c/Program Files/dotnet:/cygdrive/c/Program Files/Microsoft SQL Server/Client SDK/ODBC/170/Tools/Binn:/cygdrive/c/Program Files (x86)/IncrediBuild:/cygdrive/c/WINDOWS/system32:/cygdrive/c/WINDOWS:/cygdrive/c/WINDOWS/System32/Wbem:/cygdrive/c/WINDOWS/System32/WindowsPowerShell/v1.0:/cygdrive/c/WINDOWS/System32/OpenSSH:/cygdrive/c/Program Files (x86)/Microsoft SQL Server/150/Tools/Binn:/cygdrive/c/Program Files/Microsoft SQL Server/150/Tools/Binn:/cygdrive/c/Program Files (x86)/Microsoft SQL Server/150/DTS/Binn:/cygdrive/c/Program Files/Microsoft SQL Server/150/DTS/Binn:/cygdrive/c/Program Files (x86)/Microsoft SQL Server/Client SDK/ODBC/130/Tools/Binn:/cygdrive/c/Program Files (x86)/Microsoft SQL Server/140/Tools/Binn:/cygdrive/c/Program Files (x86)/Microsoft SQL Server/140/DTS/Binn:/cygdrive/c/Program Files (x86)/Microsoft SQL Server/140/Tools/Binn/ManagementStudio:/cygdrive/d/Git/cmd:/cygdrive/c/Users/sde26/AppData/Local/Microsoft/WindowsApps:/cygdrive/c/Users/sde26/.dotnet/tools)

which: no text2image in (./api)

which: no text2image in (./training)

ERROR: 'text2image' not found

I found text2image comes after running command 'make training'.

Can you please help me how this can be done in WIndows 10?

Regards,

Samruddhi

[Zdenko Podobny]

How did you install tesseract? Did you also install training tools?

Zdenko

po 23. 8. 2021 o 15:34 Samruddhi Dhake <sam22...@gmail.com> napísal(a):

--
You received this message because you are subscribed to the Google Groups "tesseract-ocr" group.
To unsubscribe from this group and stop receiving emails from it, send an email to tesseract-oc...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/tesseract-ocr/5adf563d-117b-4bd8-a283-dd21e53575f4n%40googlegroups.com.

[Samruddhi Dhake]

I have generated training tools through source code.

[Samruddhi Dhake]

I haven't installed Tesseract. I have kept in a folder and I am running exe by giving its path. I have generated training tools through source code.

To create box file, command->(I gave absoulute path of tesseract.exe)

..\tesseract.exe Dim4.tif Dim4 lstmbox

To create box file, command->

tesseract.exe Dim4.tif Dim4 lstm.train

To create unicharset, command->

unicharset_extractor.exe --output_unicharset ..\own.unicharset ..\langdata\eng\eng.training_text

And to create trainned data, using tesstrain.sh command,

.\src\training\tesstrain.sh --fonts_dir C:\Windows\Fonts --lang eng --linedata_only --noextract_font_properties --langdata_dir langdata --tessdata_dir tessdata --output_dir trainneddata

Regards,

Samruddhi

[Zdenko Podobny]

So there are only 2 possibilities:

1. Install tesseract and training tools
2. Learn how to handle & use not installed sw. This option is not related to tesseract.

Zdenko

ut 24. 8. 2021 o 9:17 Samruddhi Dhake <sam22...@gmail.com> napísal(a):

To view this discussion on the web visit https://groups.google.com/d/msgid/tesseract-ocr/853c21b6-9b58-42ea-929e-f9b932098bbdn%40googlegroups.com.

[Samruddhi Dhake]

How to install tesseract and training tools on Windows? 

Do I have to install Tesseract Windows exe?

Samruddhi

[Samruddhi Dhake]

Can you please provide link for steps to install Tesseract and training tools on Windows?

Samruddhi

[Samruddhi Dhake]

As you suggested, I installed Tesseract v5.0.0 on my Windows machine  (Index of /tesseract (uni-mannheim.de)). This included training tools too.

I performed all the previous steps(boxfile, lstmf file,unicharset)

But still after running tesstrain.sh command in Cygwin, I am getting following error,

$ ./src/training/tesstrain.sh --fonts_dir C:/Windows/Fonts --lang eng --linedata_only --noextract_font_properties --langdata_dir "C:/Program Files/Tesseract-OCR/langdata" --tessdata_dir "C:/Program Files/Tesseract-OCR/tessdata" --output_dir D:/Bugs/1206806/folder/trainneddata

Creating new directory D:/Bugs/1206806/folder/trainneddata

=== Starting training for language 'eng'

which: no text2image in (/usr/local/bin:/usr/bin:/cygdrive/c/Program Files/Microsoft MPI/Bin:/cygdrive/c/buildtools:/cygdrive/c/Program Files (x86)/NVIDIA Corporation/PhysX/Common:/cygdrive/c/Program Files (x86)/Intel/Intel(R) Management Engine Components/iCLS:/cygdrive/c/Program Files/Intel/Intel(R) Management Engine Components/iCLS:/cygdrive/c/Python25:/cygdrive/c/ProgramData/Oracle/Java/javapath:/cygdrive/c/Perl/site/bin:/cygdrive/c/Perl/bin:/cygdrive/c/Oracle12C_64bCli/client_1/bin:/cygdrive/c/Oracle12C_32bCli/client_1/bin:/cygdrive/c/windows/system32:/cygdrive/c/windows:/cygdrive/c/windows/System32/Wbem:/cygdrive/c/windows/System32/WindowsPowerShell/v1.0:/cygdrive/c/windows/System32/OpenSSH:/cygdrive/c/Program Files (x86)/Microsoft SQL Server/100/Tools/Binn:/cygdrive/c/Program Files/Microsoft SQL Server/100/Tools/Binn:/cygdrive/c/Program Files/Microsoft SQL Server/100/DTS/Binn:/cygdrive/c/Program Files/Microsoft/Web Platform Installer:/cygdrive/c/Program Files (x86)/Microsoft ASP.NET/ASP.NET Web Pages/v1.0:/cygdrive/c/Program Files/Microsoft SQL Server/110/Tools/Binn:/cygdrive/c/windows/system32/config/systemprofile/.dnx/bin:/cygdrive/c/Program Files/Microsoft DNX/Dnvm:/cygdrive/c/Program Files (x86)/Windows Kits/8.1/Windows Performance Toolkit:/cygdrive/c/Program Files/Microsoft SQL Server/130/Tools/Binn:/cygdrive/c/Program Files (x86)/Windows Kits/10/Windows Performance Toolkit:/cygdrive/c/Program Files (x86)/Oracle/Berkeley DB 12cR1 6.0.20/bin:/cygdrive/c/Program Files/dotnet:/cygdrive/c/Program Files/Microsoft SQL Server/Client SDK/ODBC/170/Tools/Binn:/cygdrive/c/Program Files (x86)/IncrediBuild:/cygdrive/c/WINDOWS/system32:/cygdrive/c/WINDOWS:/cygdrive/c/WINDOWS/System32/Wbem:/cygdrive/c/WINDOWS/System32/WindowsPowerShell/v1.0:/cygdrive/c/WINDOWS/System32/OpenSSH:/cygdrive/c/Program Files (x86)/Microsoft SQL Server/150/Tools/Binn:/cygdrive/c/Program Files/Microsoft SQL Server/150/Tools/Binn:/cygdrive/c/Program Files (x86)/Microsoft SQL Server/150/DTS/Binn:/cygdrive/c/Program Files/Microsoft SQL Server/150/DTS/Binn:/cygdrive/c/Program Files (x86)/Microsoft SQL Server/Client SDK/ODBC/130/Tools/Binn:/cygdrive/c/Program Files (x86)/Microsoft SQL Server/140/Tools/Binn:/cygdrive/c/Program Files (x86)/Microsoft SQL Server/140/DTS/Binn:/cygdrive/c/Program Files (x86)/Microsoft SQL Server/140/Tools/Binn/ManagementStudio:/cygdrive/d/Git/cmd:/cygdrive/c/Users/sde26/AppData/Local/Microsoft/WindowsApps:/cygdrive/c/Users/sde26/.dotnet/tools)

which: no text2image in (./api)

which: no text2image in (./training)

ERROR: 'text2image' not found

Am I missing something? Can you please guild me?

Regards,

Samruddhi

[Zdenko Podobny]

Honestly, I have no clue what you are doing: text2image is at the same location as the tesseract executable. So if you have tesseract in the path, text2image must work too. 

Zdenko

st 25. 8. 2021 o 16:26 Samruddhi Dhake <sam22...@gmail.com> napísal(a):

To view this discussion on the web visit https://groups.google.com/d/msgid/tesseract-ocr/79bf5824-5f74-4dc9-b2da-269840d1dc7fn%40googlegroups.com.

[Samruddhi Dhake]

Hi,

Text2Image error is gone. I am getting font-config error.

SDE26@DTP-SDE26-IND /cygdrive/c/Program Files/Tesseract-OCR

$ ./src/training/tesstrain.sh --fonts_dir C:/Windows/Fonts --lang eng --linedata_only --noextract_font_properties --langdata_dir "C:/Program Files/Tesseract-OCR/langdata" --tessdata_dir "C:/Program Files/Tesseract-OCR/tessdata" --output_dir D:\Test\trainneddata

Creating new directory D:Testtrainneddata

=== Starting training for language 'eng'

[Mon Aug 30 15:34:53 IST 2021] /cygdrive/c/Program Files/Tesseract-OCR/text2image --fonts_dir=C:/Windows/Fonts --ptsize 12 --font=Arial Bold --outputbase=/tmp/font_tmp.hbC9F3LEQX/sample_text.txt --text=/tmp/font_tmp.hbC9F3LEQX/sample_text.txt --fontconfig_tmpdir=/tmp/font_tmp.hbC9F3LEQX

Unable to open '/tmp/font_tmp.hbC9F3LEQX/fonts.conf' for writing

Fontconfig error: Cannot load default config file

Could not find font named 'Arial Bold'.

Please correct --font arg.

ERROR: Program Program failed. Abort.

I have Arial Bold font on my machine. Don't know why it cannot find. And in /tmp/ folder there is no font_tmp.hbC9F3LEQX where fonts.conf cannot be opened for writing.

How can I resolve this?

Regards,

Samruddhi

[Jaspreet Kaur]

Check the details of all fonts. It gives error for font arial bold. Check this font exists or not. It not then download arial bold font first.

To view this discussion on the web visit https://groups.google.com/d/msgid/tesseract-ocr/6492b2e2-060c-41a5-97bd-dfc238656cb4n%40googlegroups.com.

[Samruddhi Dhake]

Yes, I have Arial Bold on my machine.

>Fontconfig error: Cannot load default config file

And for this do I have add in environment variables? Any idea?

Regards,

Samruddhi

[Jaspreet Kaur]

Okay, then check all the that are required for your training

To view this discussion on the web visit https://groups.google.com/d/msgid/tesseract-ocr/ea28d7d1-fee2-4c38-9e83-8dc6dff63dd7n%40googlegroups.com.

[Zdenko Podobny]

First of all: use quotes for multi word names, or escape space/special symbols (e.g. --font="Arial Bold")

Next: fix error message: "Unable to open '/tmp/font_tmp.hbC9F3LEQX/fonts.conf' for writing" 

Next: check available font for text2image with option --list_available_fonts

etc...

PS: I would suggest using linux for training instead of windows (e.g. in WSL[1])

[1] https://docs.microsoft.com/en-us/windows/wsl/install-win10

Zdenko

po 30. 8. 2021 o 12:12 Samruddhi Dhake <sam22...@gmail.com> napísal(a):

To view this discussion on the web visit https://groups.google.com/d/msgid/tesseract-ocr/6492b2e2-060c-41a5-97bd-dfc238656cb4n%40googlegroups.com.

[Samruddhi Dhake]

Hi,

I am running command ->

./src/training/tesstrain.sh --fonts_dir C:/Windows/Fonts --lang eng --linedata_only --noextract_font_properties --langdata_dir "C:/Program Files/Tesseract-OCR/langdata" --tessdata_dir "C:/Program Files/Tesseract-OCR/tessdata" --output_dir D:\Test\trainneddata

And after hitting enter -> (processing)

=== Starting training for language 'eng'

[Mon Aug 30 16:51:10 IST 2021] /cygdrive/c/Program Files/Tesseract-OCR/text2image --fonts_dir=C:/Windows/Fonts/ --ptsize 12 --font=Arial Bold --outputbase=/tmp/font_tmp.s9cdSHrzKS/sample_text.txt --text=/tmp/font_tmp.s9cdSHrzKS/sample_text.txt --fontconfig_tmpdir=/tmp/font_tmp.s9cdSHrzKS

Unable to open '/tmp/font_tmp.s9cdSHrzKS/fonts.conf' for writing

Fontconfig error: Cannot load default config file

Could not find font named 'Arial Bold'.

Please correct --font arg.

ERROR: Program Program failed. Abort.

I will break it to ask few queries.

[Mon Aug 30 16:51:10 IST 2021] /cygdrive/c/Program Files/Tesseract-OCR/text2image --fonts_dir=C:/Windows/Fonts/ --ptsize 12 --font=Arial Bold --outputbase=/tmp/font_tmp.s9cdSHrzKS/sample_text.txt --text=/tmp/font_tmp.s9cdSHrzKS/sample_text.txt --fontconfig_tmpdir=/tmp/font_tmp.s9cdSHrzKS

Unable to open '/tmp/font_tmp.s9cdSHrzKS/fonts.conf' for writing

----> Here, I am not giving input as Arial Bold. Outputbase , this should create temp folder 'font_tmp.s9cdSHrzKS' but its not creating.

And so does fontconfig_tmpdir'. So it is giving writing error

Fontconfig error: Cannot load default config file

----> To resolve this error, I added FONTCONFIG_FILE=%WINDIR%\fonts.conf to environment variables(referring https://forums.wesnoth.org/viewtopic.php?t=22821) 

But still not resolved.

I was checking-> text2image.exe ----list_available_fonts

And after hitting enter, I got -> Fontconfig warning: "/tmp\fonts.conf", line 4: empty font directory name ignored

The contents of the fonts.conf file which gets created are->

<?xml version="1.0"?>

<!DOCTYPE fontconfig SYSTEM "fonts.dtd">

<fontconfig>

<dir></dir>

<cachedir>/tmp</cachedir>

<config></config>

</fontconfig>

Can you please help me how can this be resolved? Or Am I giving correct tesstrain.sh command with its args?

Regards,
Samruddhi

[Jaspreet Kaur]

Use Linux instead of windows when you are working for training tesseract.

To view this discussion on the web visit https://groups.google.com/d/msgid/tesseract-ocr/a274f441-5986-415c-a0a0-e05de6a3e790n%40googlegroups.com.

[Zdenko Podobny]

try run this:

"C:\Program Files\Tesseract-OCR\text2image.exe" --fonts_dir=%WINDIR%/Fonts --fontconfig_tmpdir=/tmp --list_available_fonts

Zdenko

po 30. 8. 2021 o 16:45 Samruddhi Dhake <sam22...@gmail.com> napísal(a):

To view this discussion on the web visit https://groups.google.com/d/msgid/tesseract-ocr/a274f441-5986-415c-a0a0-e05de6a3e790n%40googlegroups.com.

[Samruddhi Dhake]

>"C:\Program Files\Tesseract-OCR\text2image.exe" --fonts_dir=%WINDIR%/Fonts --fontconfig_tmpdir=/tmp --list_available_fonts

This worked. I got list of available fonts which contains Arial and Arial Bold too.

Now this time,in Cygwin Bash, I tried giving --fontlist 'Arial' for tesstrain.sh

Command->

$ ./src/training/tesstrain.sh --fonts_dir %WINDIR%/Fonts/ --lang eng --linedata_only --noextract_font_properties --langdata_dir 'C:/Program Files/Tesseract-OCR/langdata' --tessdata_dir 'C:/Program Files/Tesseract-OCR/tessdata' --output_dir D:/Test/trainneddata --fontlist 'Arial'

=== Starting training for language 'eng'

[Tue Aug 31 19:19:05 IST 2021] /cygdrive/c/Program Files/Tesseract-OCR/text2image --fonts_dir=%WINDIR%/Fonts/ --ptsize 12 --font=Arial --outputbase=/tmp/font_tmp.0doGBqWc3I/sample_text.txt --text=/tmp/font_tmp.0doGBqWc3I/sample_text.txt --fontconfig_tmpdir=/tmp/font_tmp.0doGBqWc3I

Unable to open '/tmp/font_tmp.0doGBqWc3I/fonts.conf' for writing

Fontconfig error: Cannot load default config file

Could not find font named 'Arial'.

Please correct --font arg.

ERROR: Program Program failed. Abort.

Still I am getting this font.conf error. Any idea how to resolve this font.conf error?

Regards,

Samruddhi

[Samruddhi Dhake]

In this text2image, there is an rgument --fontconfig_tempdir which creates temp folder where fonts.conf gets added.

I checked /tmp/, no other tempfolder is created( font_tmp.0doGBqWc3I)

Has anybody this issue?

Regards,

Samruddhi

[Jaspreet Kaur]

Tesstrain.sh work for you ?

To view this discussion on the web visit https://groups.google.com/d/msgid/tesseract-ocr/4d0f22e4-cc3f-4487-a024-363e79ad8598n%40googlegroups.com.

[Samruddhi Dhake]

No. Tessstrain.sh didn't work. I am running tesstrain.sh on cygwin.

 Command->

$ ./src/training/tesstrain.sh --fonts_dir %WINDIR%/Fonts/ --lang eng --linedata_only --noextract_font_properties --langdata_dir 'C:/Program Files/Tesseract-OCR/langdata' --tessdata_dir 'C:/Program Files/Tesseract-OCR/tessdata' --output_dir D:/Test/trainneddata --fontlist 'Arial'

After hitting enter for tesstrain.sh, it is processing text2image and giving following error

=== Starting training for language 'eng'

[Tue Aug 31 19:19:05 IST 2021] /cygdrive/c/Program Files/Tesseract-OCR/text2image --fonts_dir=%WINDIR%/Fonts/ --ptsize 12 --font=Arial --outputbase=/tmp/font_tmp.0doGBqWc3I/sample_text.txt --text=/tmp/font_tmp.0doGBqWc3I/sample_text.txt --fontconfig_tmpdir=/tmp/font_tmp.0doGBqWc3I

Unable to open '/tmp/font_tmp.0doGBqWc3I/fonts.conf' for writing

Fontconfig error: Cannot load default config file

Could not find font named 'Arial'.

Please correct --font arg.

ERROR: Program Program failed. Abort.

As per previous suggestions, I ran text2image.exe command on cmd and its working and giving me all available fonts.

Then after running tesstrain.sh, why text2image command is failing and it is not creating tempfolder under /tmp/ and I am getting fonts.config error.

It is expected that fonts.config file which gets created in tempfolder(here in my case font_tmp.0doGBqWc3I) should gets written and it should include font 'Arial' and then Arial font can be found.

Don't why it is not creating..

Regards,

Samruddhi

[Jaspreet Kaur]

Okay, 

Wait you are working for English language right?

What kind of dataset you used here.

To view this discussion on the web visit https://groups.google.com/d/msgid/tesseract-ocr/3fbe32ef-5477-42c4-911b-b980b24cea9cn%40googlegroups.com.

[Samruddhi Dhake]

Yes, I am working for eng language.

I am using tessdata.(C:\Program Files\Tesseract-OCR\tessdata)

[Jaspreet Kaur]

I mean working with font only?  

Or images??

To view this discussion on the web visit https://groups.google.com/d/msgid/tesseract-ocr/595017f3-630a-4707-b4b3-a5aeed9e7a53n%40googlegroups.com.

[Samruddhi Dhake]

For images.

I have to create my own trainneddata for my images. So for that I am following steps mentioned in this documentation https://tesseract-ocr.github.io/tessdoc/tess4/TrainingTesseract-4.00.html

As per the steps I have created box file, lstm file and unicharset file. And next step is to create traineddata using tesstrain.sh followed by the next step i.e. lstmtraining.exe .

I am getting such errors while performing at step tesstrain.sh.

[Jaspreet Kaur]

Check this 

https://github.com/tesseract-ocr/tesseract/issues/1685

To view this discussion on the web visit https://groups.google.com/d/msgid/tesseract-ocr/89197941-16d3-4747-b280-95ddb9979b40n%40googlegroups.com.

[Samruddhi Dhake]

I checked this link. It has no tesstrain reference.

Tesstrain internally calls text2image.exe.

So if I ran text2image.exe, how will I get trainned data? What are the further steps to get trainneddata?

[Samruddhi Dhake]

Hi,

Though I am stuck with tesstrain.sh command, I have one question related to lstmtraining.exe command.

training/lstmtraining --debug_interval 100 \ --traineddata ~/tesstutorial/engtrain/eng/eng.traineddata \ --net_spec '[1,36,0,1 Ct3,3,16 Mp3,3 Lfys48 Lfx96 Lrx96 Lfx256 O1c111]' \ --model_output ~/tesstutorial/engoutput/base --learning_rate 20e-4 \ --train_listfile ~/tesstutorial/engtrain/eng.training_files.txt \ --eval_listfile ~/tesstutorial/engeval/eng.training_files.txt \ --max_iterations 5000 &>~/tesstutorial/engoutput/basetrain.log

How do I get --train_listfile engtrain/eng.training_files.txt and --eval_listfile ~/tesstutorial/engeval/eng.training_files.txt?

Regards,

Samruddhi

[Samruddhi Dhake]