box.train completes with no errors but does not create .tr output

[danageva]

Hi,
I'm running 3.05.00dev on Ubuntu 14.04 LTS.
When running:
tesseract eng.Arial.exp0.tif eng.Arial.exp0 box.train
I'm getting a simple one-line output:
Tesseract Open Source OCR Engine v3.05.00dev with Leptonica

However, no .tr output file is created (anywhere in the filesystem).

My work dir listing is:

Arial.ttf
common.punc
eng.Arial.exp0.box
eng.Arial.exp0.tif
training-text.txt

Running with gdb doesn't give anything additional.

Anything I can look for for extra info? Any ideas what might be causing this?

Thanks

[rodrigosalinas]

I'm having exactly the same problem than danageva with the same version but over OS X. Apparently zdenop closed the issue with a deletion in two files. I don't have those lines in my installed version. Can you help me please?

[amitdo]

@rodrigosalinas
The issue was solved (again).
Try latest code from the repo.

[aiwaz]

I'm having exactly the same problem. I have the latest code from the repo.
Edit: It works with v3.04.00

[amitdo]

You should provide more details.

I have the latest code from the repo.

Which commit exactly, 1826ac1 ?

What OS do you use and the specific version of this OS. For example: Ubuntu 14.04, OS X 10.11, Windows 10.

Also, please provide the exact command you used and attach as zip file the tif and box files.

[aiwaz]

git rev-parse HEAD command gives me 1826ac1

uname -a
Linux ubuntu 3.19.0-42-generic #48-Ubuntu SMP Thu Dec 17 22:54:45 UTC 2015 x86_64 x86_64 x86_64 GNU/Linux
lsb_release -a
Distributor ID: Ubuntu Description: Ubuntu 15.04 Release: 15.04 Codename: vivid

The exact command is:
tesseract ult.dejavu.exp0.tif ult.dejavu.exp0 box.train
The output:
Tesseract Open Source OCR Engine v3.05.00dev with Leptonica Segmentation fault (core dumped)
If I use it with sudo:
sudo tesseract ult.dejavu.exp0.tif ult.dejavu.exp0 box.train
The output:
Tesseract Open Source OCR Engine v3.05.00dev with Leptonica

Required files attached.
ult.zip

[amitdo]

I have Ubuntu 14.04 64 bit.

Commit 1826ac1.

tesseract ult.dejavu.exp0.tif ult.dejavu.exp0 box.train

Output:

Tesseract Open Source OCR Engine v3.05.00dev with Leptonica
Page 1
APPLY_BOXES:
Boxes read from boxfile: 1360
Found 1360 good blobs.
Generated training data for 292 words
Warning in pixReadMemTiff: tiff page 1 not found

ult.dejavu.exp0.tr.zip

My guess is that you didn't install tesseract properly.

Did you run all these commands last time you installed tesseract?

./autogen.sh
./configure --disable-cube
make
sudo make install
sudo ldconfig
make training
sudo make training-install

[aiwaz]

I installed it as:

./autogen.sh
./configure --prefix=/home/username/tessbin
make
make install
make training
make training-install

[amitdo]

if you have previously installed tesseract system-wide (for example, under /usr/local) the command 'tesseract ...' will use that system-wide executable.

[amitdo]

Try this command to see if you have another installation of Tesseract in your machine.

find /usr -name "tesseract"

[stweil]

... or call /home/username/tessbin/bin/tesseract directly.

[aiwaz]

Nope, there's no other tesseract in my machine. find does not return
anything.

On 27 January 2016 at 18:36, Amit Dovev [email protected] wrote:

Try this command to see if you have another installation of Tesseract in
your machine.

find /usr -name "tesseract"

—
Reply to this email directly or view it on GitHub
#64 (comment)
.

[stweil]

@aiwaz, could you try this command:

gdb --args tesseract ult.dejavu.exp0.tif ult.dejavu.exp0 box.train

(Enter run on the gdb command line, and when it reports an error info stack)

[aiwaz]

gdb --args tesseract ult.dejavu.exp0.tif ult.dejavu.exp0 box.train
GNU gdb (Ubuntu 7.9-1ubuntu1) 7.9
Copyright (C) 2015 Free Software Foundation, Inc.
License GPLv3+: GNU GPL version 3 or later <http://gnu.org/licenses/gpl.html>
This is free software: you are free to change and redistribute it.
There is NO WARRANTY, to the extent permitted by law.  Type "show copying"
and "show warranty" for details.
This GDB was configured as "x86_64-linux-gnu".
Type "show configuration" for configuration details.
For bug reporting instructions, please see:
<http://www.gnu.org/software/gdb/bugs/>.
Find the GDB manual and other documentation resources online at:
<http://www.gnu.org/software/gdb/documentation/>.
For help, type "help".
Type "apropos word" to search for commands related to "word"...
Reading symbols from ../../bin/tesseract...done.
(gdb) run
Starting program: /home/azukausk/tessbin/bin/tesseract ult.dejavu.exp0.tif ult.dejavu.exp0 box.train
[Thread debugging using libthread_db enabled]
Using host libthread_db library "/lib/x86_64-linux-gnu/libthread_db.so.1".
Tesseract Open Source OCR Engine v3.05.00dev with Leptonica

Program received signal SIGSEGV, Segmentation fault.
tesseract::TessResultRenderer::BeginDocument (this=this@entry=0x83f0b0, title=title@entry=0x7ffff78a8fea "")
    at renderer.cpp:57
57    bool ok = BeginDocumentHandler();
(gdb) info stack
#0  tesseract::TessResultRenderer::BeginDocument (this=this@entry=0x83f0b0, title=title@entry=0x7ffff78a8fea "")
    at renderer.cpp:57
#1  0x00007ffff76cee2f in tesseract::TessBaseAPI::ProcessPagesInternal (this=this@entry=0x7fffffffe2a0, 
    filename=<optimized out>, retry_config=retry_config@entry=0x0, timeout_millisec=timeout_millisec@entry=0, 
    renderer=0x83f0b0) at baseapi.cpp:1166
#2  0x00007ffff76cf570 in tesseract::TessBaseAPI::ProcessPages (this=this@entry=0x7fffffffe2a0, 
    filename=<optimized out>, retry_config=retry_config@entry=0x0, timeout_millisec=timeout_millisec@entry=0, 
    renderer=<optimized out>) at baseapi.cpp:1074
#3  0x0000000000401f5c in main (argc=<optimized out>, argv=0x7fffffffe448) at tesseractmain.cpp:429
(gdb)

[aiwaz]

@stweil, I always call tesseract directly. It resides in /home/username/tessbin/bin/tesseract , just as you wrote. I created an alias for it in bashrc.

[amitdo]

What's the output of:

printenv TESSDATA_PREFIX

tesseract --list-langs

[aiwaz]

TESSDATA_PREFIX is empty.

tesseract --list-langs
List of available languages (1):
eng

[amitdo]

tesseract ult.dejavu.exp0.tif ult txt hocr

Does this command produce ult.txt and ult.hocr ?

Maybe you should try to set TESSDATA_PREFIX environment variable or use --tessdata-dir parameter.

[aiwaz]

tesseract ult.dejavu.exp0.tif ult txt hocr
Tesseract Open Source OCR Engine v3.05.00dev with Leptonica
Page 1
Error opening data file /home/azukausk/tessbin/share/tessdata/osd.traineddata
Please make sure the TESSDATA_PREFIX environment variable is set to the parent directory of your "tessdata" directory.
Failed loading language 'osd'
Tesseract couldn't load any languages!
Warning: Auto orientation and script detection requested, but osd language failed to load
Warning in pixReadMemTiff: tiff page 1 not found

Why it wants to load the "osd" language?

The files are produced:

ll
total 21500
drwxrwxr-x 2 azukausk azukausk     4096 Jan 27 13:49 configs
-rw-rw-r-- 1 azukausk azukausk 21876550 Jan 25 15:41 eng.traineddata
-rw-r--r-- 1 azukausk azukausk      568 Jan 27 13:49 pdf.ttf
drwxrwxr-x 2 azukausk azukausk     4096 Jan 27 13:49 tessconfigs
-rw-rw-r-- 1 azukausk azukausk     1649 Jan 25 15:56 training_text
-rw-rw-r-- 1 azukausk azukausk    38706 Jan 27 12:53 ult.dejavu.exp0.box
-rw-rw-r-- 1 azukausk azukausk    39984 Jan 27 12:53 ult.dejavu.exp0.tif
-rw-rw-r-- 1 azukausk azukausk    36383 Jan 28 12:03 ult.hocr
-rw-rw-r-- 1 azukausk azukausk     1685 Jan 28 12:03 ult.txt

When I set the TESSDATA_PREFIX variable the output is the same:

TESSDATA_PREFIX=/home/azukausk/tessbin/share/tessdata
echo $TESSDATA_PREFIX
/home/azukausk/tessbin/share/tessdata
tesseract ult.dejavu.exp0.tif ult txt hocr
Tesseract Open Source OCR Engine v3.05.00dev with Leptonica
Page 1
Error opening data file /home/azukausk/tessbin/share/tessdata/osd.traineddata
Please make sure the TESSDATA_PREFIX environment variable is set to the parent directory of your "tessdata" directory.
Failed loading language 'osd'
Tesseract couldn't load any languages!
Warning: Auto orientation and script detection requested, but osd language failed to load
Warning in pixReadMemTiff: tiff page 1 not found

And one more question: when I configure tesseract with --prefix option, I expect tesseract to be smart enough to know where my data are. Why do I have to set TESSDATA_PREFIX variable explicitly?

[stweil]

Setting TESSDATA_PREFIX is not needed as long as your tessdata directory is at the right place ($PREFIX/share/tessdata).

osd.traineddata will be used for auto orientation and script detection, no matter which language you have selected.

[amitdo]

OK, I was able to reproduce the issue.

I will try to figure out why this is happening when installing tesseract in
/home/myusername/tessbin

[amitdo]

I solved @aiwaz issue. More info will follow later.

[amitdo]

@aiwaz, please test my fix.

[amitdo]

zdenko merged my fix to master.

[aiwaz]

I tested the fix and confirm that tesseract ult.dejavu.exp0.tif ult.dejavu.exp0 box.train now produces .tr file.
Thank you guys.

[amitdo]

👍
Thanks for reporting the issue.