Tesseract lang creation script

Tesseract lang creation script

#!/bin/bash

TRAININGDIR="./Trained"
SUFFIX="tif"
FONTPROPERTIES="$TRAININGDIR/font_properties"

for f in $(find $TRAININGDIR -iname "*.$SUFFIX")
do
    # Get the name of the file
    filename=$(echo ${f%%.$SUFFIX} | sed 's#^.*/##')
    
    # Get the path of the file
    path=`dirname $f`
    
    # Train the files
    tesseract $f $path/$filename nobatch box.train
    
    # While we are at it, create a null font_properties
    echo $filename 0 0 0 0 0 >> $FONTPROPERTIES
    
done

# Get a list of all the Box files
ALLBOX=""
for f in $(find $TRAININGDIR -iname "*.box")
do
    ALLBOX=${ALLBOX}" "${f} 
done

# Create the unicharset
unicharset_extractor $ALLBOX
# Move it to the appropriate place
mv unicharset $TRAININGDIR

# Get a list of all the Box files
ALLTR=""
for f in $(find $TRAININGDIR -iname "*.tr")
do
    ALLTR=${ALLTR}" "${f} 
done

# Perform the clustered training
mftraining -F $FONTPROPERTIES -U $TRAININGDIR/unicharset $ALLTR

cntraining $ALLTR

# Move files to appropriate place
mv inttemp $TRAININGDIR/blast.inttemp
mv mfunicharset $TRAININGDIR/blast.mfunicharset
mv Microfeat $TRAININGDIR/blast.Microfeat
mv normproto $TRAININGDIR/blast.normproto
mv pffmtable $TRAININGDIR/blast.pffmtable
mv $TRAININGDIR/unicharset $TRAININGDIR/blast.unicharset

combine_tessdata $TRAININGDIR/blast.