Hi,
Finally I found how can we add RTl support to 3.02
I Wrote a code in python. it can convert unicharset file to RTL supporting
one
After making unicharset file you should run attached code to correcting
wrong properties.
I attached this code may be it will useful for other RTL languages!
For running this code you should installed python on your PC
your,
Reza
On Thursday, October 18, 2012 11:38:39 PM UTC+2, Reza M wrote:
>
> Hi,
> I made a simple traineddata for Persian it recognized characters but it
> changes words directions for example instead of رضا it writes اضر
>
> I tried to use Data of Herby or Arabic but it doesn't work
> I tried 3.02 and it doesn't work correctly for my data!
>
> do you know how can i make my data like Arabic with cub mode? or like
> Herby that works correctly for RTL?
> there is many languages that they are RTL would you please tell us how did
> you made Arabic file?
>
> yours,
> reza
>
--
You received this message because you are subscribed to the Google
Groups "tesseract-ocr" group.
To post to this group, send email to [email protected]
To unsubscribe from this group, send email to
[email protected]
For more options, visit this group at
http://groups.google.com/group/tesseract-ocr?hl=en
?#!/usr/bin/python
# -*- coding: utf-8 -*-
# Reza1615
# Distributed under the terms of the CC-BY-SA 3.0 .
import codecs,re
filesample = 'unicharset'
lang=u'Arabic' # for Persian also uses Arabic
lines=u'\n'
couple={}
couple_ch={u"<":u">",u">":u"<",u"(":u")",u")":u"(",u"[":u"]",u"]":u"[",u"{":u"}",u"}":u"{",u"?":u"?",u"?":u"?"}
text = codecs.open( filesample,'r' ,'utf8' )
text = text.read()
for line in text.split(u'\n'):
column=line.split(u' ')
count=-1
new_line=u' '
for item in column: # Changing items in Columns
count+=1
if count==3:
if column[0].strip() in u"~#@?'\/<>()[]}{.,?!$*+-_&0123456789????????????????????????????":#Common charcters
new_line+=item.replace(u'NULL',u'Common ')
else:
new_line+=item.replace(u'NULL',lang+u' ')
continue
if count==5:
if column[0].strip() in u"0123456789":# English numbers
new_line+=u'2 '
elif column[0].strip() in u"+-":
new_line+=u'3 '
elif column[0].strip() in u"#$?":# Dollar sign and it's alternatives
new_line+=u'4 '
elif column[0].strip() in u"????????????????????":# Your language's numbers and its alternatives
new_line+=u'5 '
elif column[0].strip() in u"/.,??":# Your language's separators
new_line+=u'6 '
elif column[0].strip() in u"~@?'\<>()[]}{?!*_&??":
new_line+=u'10 '
else:
new_line+=u'13 '
continue
if count==6:
if column[0].strip() in u"<>()[]}{??":# Pair Characters
couple[column[0].strip()]=column[4].strip()
new_line+=u'[$['+column[0].strip()+u']$] '
else:
new_line+=column[4].strip()+u' '
continue
new_line+=item+u' '
lines+=new_line.strip()+u'\n'
# Changing paired charcters column number
total_couples=re.findall(ur"\[\$\[.*?\]\$\]",lines, re.S)
if total_couples:
for i in total_couples:
i=i.replace(u'[$[',u'').replace(u']$]',u'').strip()
try:
lines=lines.replace(u'[$['+i+u']$]',couple[couple_ch[i]])
except:
lines=lines.replace(u'[$['+i+u']$]',couple[i])
print i + u' is not paired! please add '+couple_ch[i] +u' to your Box'
else:
print "It couldn't find couple characters like <>()[]}{?? "
# Work is finished!
print "Now your unicharset is set for RTL languages (Arnew_lineic, Persian)!"
with codecs.open(filesample ,mode = 'w',encoding = 'utf8' ) as f:
f.write( lines.strip())
