Updated 2015-08-22 10:23:08 by linuxpeter

This is a working solution

هذا حل فعلا يعمل و هو سهل الاستعمال

The following procedure render_arabic does not completely solve the rendering the Arabic characters to be fully connected. However, it does reach the wanted target and goal of making the Arabic smoothly readable. So you can now type in the code as such:
set arabic_rendered_sentence [render_arabic "مرحبا بكم"]

Then get in your Tk something very near to the following result.

Please just understand I am using the text to explain and imagine it with the spaces are just only one millimeter wide.
مـ ر حـ بـا بـ كم

Download the script here OR here

The video, Render Arabic Language Words in TCL and Tk (Video Take 2) , This video on YouTube explains how it works from within the code.

If you are planning to do inputs, then do the following steps:

  • Use a text widget for your input.
  • Include to the start of your code the two the procedures mentioned below: render_arabic $arabic_string and text_binding_for_rtl $text_widget_path $k
  • Of course, you would be using a number of text widgets. Therefore, make a list of your text widget windows paths like the following:
set list_of_widget_window_paths [list .mytext1 .anotherText2 .toplevel.toplevel_text]

  • Then bind all of the text widgets as follows:
foreach text_widget_window_path $list_of_widget_window_paths {
    bind $text_widget_window_path <Key> [list text_binding_for_rtl %W %k]
}

And you're done!

Render Arabic Language Words in TCL and Tk (Video Take 3) shows it in action.

The Required Procedures  edit

render_arabic $arabic_string

makes Arabic readable when displayed in a Tk widget.
#! /usr/bin/wish

#Code written by Rani Fayez Ahmad (Superlinux)
#Website: http://www.superlinux.net

#The following procedure is used to extract all ASCII string parts from the Unicode string. in tk_messageBox after rendering to Arabic they come reversed
#So this fix has been added
# It will give the pairs :  (ASCII string part | ASCII string starting index in the Unicode string) .
proc list_of_all_ascii_parts_a_unicode_string { arabic_string} {

set ascii_parts_list [list]
set length [string length $arabic_string]
    for {set i 0} {$i< $length} {incr i} {
        set start_of_ascii $i
        
        set end_of_ascii  $start_of_ascii 
        while {[string is ascii [
            string range $arabic_string $start_of_ascii $end_of_ascii]] ==  1
            && $i<$length}  {
          
                puts [
                    string range $arabic_string $start_of_ascii $end_of_ascii]
                incr i
                incr end_of_ascii 
        }
        
        incr end_of_ascii -1
        
        set ascii_part [
            string range $arabic_string $start_of_ascii $end_of_ascii]
        
        if {[string trim $ascii_part] ne {}} {
          set ascii_parts_list [
              linsert $ascii_parts_list end [list $ascii_part $start_of_ascii]]
        }
    }
    return $ascii_parts_list
}

#a procedure to make Arabic readable when displayed in a Tk widget.
    
proc render_arabic args {
    set  arabic_string [lindex $args 0]
    set  is_messageBox [lindex $args 1]

    #The given of the problem is an Arabic sentence
    
    #Break the sentence into words
    set  words [split [string trim $arabic_string]]
    
    #Display the sentence the way TCL receives it
    #The problem is:
    #Tcl receives the Arabic letters: (i) in the reverse order (ii)
    #disconnected.  We want to re-render the Arabic to be displayed correctly
    #tk_messageBox -message $words
    
    #$count is the word index in the arabic sentence   
    set count 0
    
    #the following is just an example of how to get an arabic character index
    #number in the unicode character charts
    #set z {} ; foreach el [split ل {}] {puts [scan $el %c]}
     
    #foreach word in the arabic sentence 
    foreach word $words {
        if {[string is ascii $word]} {
            incr count
            continue
        } 
         
        #else {
        #    set splits [split $word "!@#$%^&*()_+-=~`123456790/\\"]
        #    if {[llength $splits] > 1} {
        #        set split_counter 0
        #        foreach splitting $splits {
        #            set splitting [render_arabic $splitting]
        #            lset splits $split_counter $splitting
        #            incr split_counter
        #        }
        #        set word [join splits]
        #        incr count
        #        continue
        #    }
        #}
         
        #1-get the  substring in the word without the last letter
        #we will deal with the connection of the last letter later
        set original_word $word
        set sub_word [string range $word 0 end-1]
        
        #All the letters from baa2 to yaa2 when they are NOT the last letter;
        #TCL initially has and reads them in their isolated form as in ل م س;
        #they must be converted into their initial form e.g ل م س
        #so replace and convert every occurrence of each of such letters

        #Also other Arabic-like characters like Urdu, Persian, Kurdish... etc, 
        #You may add them similarly over here
        
        set sub_word [ string map {\u0628 \ufe91} $sub_word] ;#ba2
        set sub_word [ string map {\u062A \ufe97} $sub_word] ;#Ta2
        set sub_word [ string map {\u062B \ufe9b} $sub_word] ;#thaa2
        set sub_word [ string map {\u062C \ufe9f} $sub_word] ;#Jeem
        set sub_word [ string map {\u062d \ufea3} $sub_word] ;#7aa2
        set sub_word [ string map {\u062e \ufeA7} $sub_word] ;#5aa2
        set sub_word [ string map {\u0633 \ufeb3} $sub_word] ;#seen
        set sub_word [ string map {\u0634 \ufeb7} $sub_word] ;#sheen
        set sub_word [ string map {\u0635 \ufebb} $sub_word] ;#SSaad
        set sub_word [ string map {\u0636 \ufebf} $sub_word] ;#DDhahd
        set sub_word [ string map {\u0637 \ufec3} $sub_word] ;#TTaa2
        set sub_word [ string map {\u0638 \ufec7} $sub_word] ;#tthaa2 Zah
        set sub_word [ string map {\u0639 \ufeCb} $sub_word] ;#3eyn
        set sub_word [ string map {\u063A \ufeCF} $sub_word] ;#ghyn
        set sub_word [ string map {\u0641 \ufeD3} $sub_word] ;#faa2
        set sub_word [ string map {\u0642 \ufeD7} $sub_word] ;#quaaf
        set sub_word [ string map {\u0643 \ufeDb} $sub_word] ;#kaaf
        set sub_word [ string map {\u0644 \ufedf} $sub_word] ;#lam
        set sub_word [ string map {\u0645 \ufee3} $sub_word] ;#meem
        set sub_word [ string map {\u0646 \ufee7} $sub_word] ;#noon
        set sub_word [ string map {\u0647 \ufeeb} $sub_word] ;#haa2
        set sub_word [ string map {\u064A \ufef3} $sub_word] ;#yaa2
        set sub_word [ string map {\u0626 \ufe8b} $sub_word] ;#hamza 3ala nabera (initial form of yaa2)
        
        #now replace the whole part of the word that excludes the last letter
        #with the conversion done above
        
        set word [string replace $word 0 end-1 $sub_word]
        
        #The following list of characters are the characters initial form
        #mentioned above + the tatweel chacracter
        set initials [list \u0640 \ufe90 \ufe97 \ufe9b \ufe9f \ufea3 \ufeA7 \
            \ufb3 \ufeb7 \ufebb \ufebf \ufec3 \ufec7 \ufeCb \ufeCF \ufeD3 \
            \ufeD7 \ufeDb \ufedf \ufee3 \ufee7 \ufeeb \ufef3]
        
     
        #find the character before the last.
        
        set before_last_char [string index $word end-1]
        
        #for debugging purposes just print the character before the last.
        ## puts $before_last_char
        
        #and try to see if  the character before the last is a word in the list
        #$initials defined in the previous line.
        #and if its true, then convert the last character to it's final linked
        #form
        #this way they will be joined
        if {[lsearch -ascii -inline $initials $before_last_char]
            eq $before_last_char} {
            
            #now get also last chacracter
            set last_character [string index $word end]
            
            #print it for debugging purposes
            ##puts $last_character
            
            #just to make sure that we we are matching correctly print the unicode
            #index number of the character
            ##puts [scan $last_character %c]
            if {[string is ascii $last_character]} {
                set before_last_char [render_arabic $before_last_char]
            }
            
            #\u0627 {
            #    #aleph
            #    set word [ string replace $word end end \ufe8e ]
            #}
            #now convert the last character into its final linked form
            switch -- $last_character {
                \u0628 {
                    #baa2
                    set word [string replace $word end end \ufe90]
                }
                \u0629 {
                    #taa2 marbootta
                    set word [string replace $word end end \ufe94]
                }
                \u062A {
                    #ta2 maftoo7a
                    set word [string replace $word end end \ufe96]
                }
                \u062B {
                    #thaa2
                    set word [string replace $word end end \ufe9A]
                }
                \u062c {
                    #jeem
                    set word [string replace $word end end \ufe9e]
                    puts $word
                }
                \u062d {
                    #7aa2
                    set word [string replace $word end end \ufeA2]
                }

                \u062e {
                    #5aa2
                    set word [string replace $word end end \ufea6]
                }

                \u062f {
                    #dal
                    set word [string replace $word end end \ufeaa]
                }

                \u0630 {
                    #tthal
                    set word [string replace $word end end \ufeac]
                }
                \u0631 {
                    #raa2
                    set word [string replace $word end end \ufeae]
                }
                \u0632 {
                    #zyn
                    set word [string replace $word end end \ufeaf]
                }

                \u0633 {
                    #seen
                    set word [string replace $word end end \ufeb2]
                }
                \u0634 {
                    #sheen
                    set word [string replace $word end end \ufeb6]
                }
                \u0635 {
                    #ssaad
                    set word [string replace $word end end \ufeba]
                }
                \u0636 {
                    #ddaad
                    set word [string replace $word end end \ufebe]
                }
                \u0637 {
                    #ttaa2
                    set word [string replace $word end end \ufec2]
                }
                \u0638 {
                    #tthaa2
                    set word [string replace $word end end \ufec8]
                }
                \u0639 {
                    #3ayn
                    set word [string replace $word end end \ufeca]
                }
                \u063a {
                    #ghyn
                    set word [string replace $word end end \ufece]
                }
                \u0641 {
                    #faa2
                    set word [string replace $word end end \ufed2]
                }
                \u0642 {
                    #quaaf
                    set word [string replace $word end end \ufed6]
                }
                \u0643 {
                    #kaaf
                    set word [string replace $word end end \ufeda]
                }
                  \u0644 {
                    #laam
                    set word [ string replace $word end end \ufede ]
                }
                \u0645 {
                    #meem
                    set word [string replace $word end end \ufee2]
                }
                \u0646 {
                    #noon
                    set word [string replace $word end end \ufee6]
                }
                \u0647 {
                    #haa2
                    set word [string replace $word end end \ufeea]
                }
                \u0648 {
                    #waaw
                    set word [string replace $word end end \ufeee]
                }
                \u0624 {
                    #waaw with hamza above
                    set word [ string replace $word end end \ufe86]
                }
                \u0649 {
                    #alef maqsura
                    set word [string replace $word end end \ufef0]
                }
                \u064a {
                    #yaa2
                    set word [string replace $word end end \ufef1]
                }
                default {
                    #default is nothing to do
                }
            }
     
        }
        # end of if the character before the last is a member of the list
        # $initials
         
        #now reverse every occurrence of the word for correct displaying on the
        #screen

        set arabic_string [
            regsub -all "\\m$original_word\\M" $arabic_string $word]

        #add and replace the corrected/conversion-of word with malformed one. in
        #the arabic sentence
        #the whole words in the sentence yet are still in the reverse order
        #lset words $count $word
        
        #move to the  next word
        incr count
    }
    
    #The following 2 line is left for you to see the final result. just remove
    #the comment sign (#)
    #tk_messageBox -message $words
    #puts "before return: $arabic_string \n is_messageBox=$is_messageBox"
    
    #reverse the whole string
    set arabic_string [string reverse $arabic_string]
     
    #If you see that the ASCII string parts of the whole Arabic/Unicode are
    #reversed, then add another one and only one additional parameter to the
    #Arabic/Unicode string and set it only to
    #"1" (the number ONE).

    if { $is_messageBox ==1 } {
    foreach part [list_of_all_ascii_parts_a_unicode_string $arabic_string] {
        set part_string [string reverse [ lindex $part 0 ]]
        set start_of_ascii [ lindex $part 1 ]
        set length_part_string [string length $part_string]
        
        set arabic_string [string replace $arabic_string $start_of_ascii [expr $start_of_ascii + $length_part_string -1] $part_string]
      
    }
  }
  return $arabic_string 
}

Usage  edit

An example of how to call render_arabic $arabic_string $is_messageBox:
#The Arabic sentence in the variable $msg below means in English : The program
#is not available or changeable (it should be here the word "NOT" (غير) instead
#of "CHANGEABLE" (متغير) but for debugging purposes there's an additional
#character) exists/existed 

set msg [
    render_arabic "الـبـرنـامـج غـيـر مـتـوفـر او مـتـغـيـر مـوجـود حـالئـيـا"]
puts $msg
#If you see that the ASCII string parts of the whole Arabic/Unicode are
#reversed, then add another one and only one additional parameter to the
#Arabic/Unicode string and set it only to #"1" (the number ONE).
#It might not appear to be a second parameter here, but just as you type you'll know it's a second parameter. Just test it and see for yourself
set msg [
    render_arabic "الـبـرنـامـج غـيـر مـتـوفـر او مـتـغـيـر مـوجـود حـالئـيـا" 1]
tk_messageBox -detail $msg

text_binding_for_rtl $text_widget_path $k

This procedure will eventually call render_arabic $arabic_string .
proc text_binding_for_rtl {text_widget_path k} {
    #Caution!!! *DO NOT* delete the next line of [set event_counter 0] . 
    set event_counter 0
    proc local_text_binding_for_rtl {text_widget_path k} {

        #Algorithm:
        #%k is the keycode number (%k is an integer not a hexadecimal) of the
        #pressed key.
        #we print it for correct matching and selection of the keys being
        #pressed.
        #If the %k is neither the spacebar or the Enter don't process anthing.
        #We only want to process the very last word.
        #And the Enter or the Spacebar are the triggers and the signals just to
        #say we have a new word being written in the entry.
        #Therefore, exctract the last word. The last word must be checked to
        #see whether it's completely an ASCII(Latin) string. 
        #If it's, then don't touch it and skip to the next word
        #Always keep the latin/ASCII words as they are.
        #Else, process the word as an arabic word using the procedure
        #[render_arabic] defined above
        #Also map the last word to the rest of the text widget string just to save time
        
        puts "k= $k"
        global event_counter
        #Only Key-Enter  and Key-Space ban allow the processing of the last word
        if {$k != 65 && $k != 36} {
            set event_counter 0
            return 
        } else {
            incr event_counter
        }
        
        #get the whole text as one string
        #check whether it's an [entry] widget or if it's an [text] widget
        set text_widget_type [winfo class $text_widget_path]
        set all_text ""
        switch -- $text_widget_type {
            Text {
                set all_text [$text_widget_path get 1.0 end] 
            }
            Entry {
                set all_text [$text_widget_path get ] 
            }
        }
        
        #if the text is empty or full of so many spaces, then return doing
        #nothing
        set trimmed_text [string trim $all_text]
        if {$trimmed_text == {}} {    
            return
        }
        
        if {$event_counter == 1} {
            #after trimming white spaces from both ends of the whole text,
            #convert the text into a list of words to find the last word in the
            #text set words [split [string trim $all_text] ]
            
            #for the sake of debugging and monitoring, print the list of words.
            puts $words
            
            #get the last word
            set last_word [lindex $words end]
            
            #print the last word for debugging.
            puts $last_word
            
            #the word is completely ASCII, skip to the next word by quitting this
            #event
            if {[string is ascii $last_word]} {
                return
            }  
            
            #we render the last word, which is supposed to be an arabic word
            set last_word_after_rendering [render_arabic $last_word] 
              
            #print the last word after rendering for debugging
            puts $last_word_after_rendering
            
            #replace every occurence of the original last word with the last word
            #after rendering in all the text
            #but because it everything is already rendered but the last word, it
            #will be only a replacement of the last word with 
            #the last_word_after_rendering
            set all_text [
                string map [
                    split "$last_word $last_word_after_rendering"] $all_text]
            set all_text "[string trimright $all_text] "
            #and finally re-assign the text again to the entry or text widget
            switch -- $text_widget_type {
                Text {
                    $text_widget_path replace 1.0 end $all_text 
                }
                Entry {
                    $text_widget_path delete 0 end
                    $text_widget_path insert end $all_text
                }   
            }
        }
    }
    local_text_binding_for_rtl $text_widget_path $k
}

Usage

An example of how to call text_binding_for_rtl $text_widget_path $k :
entry .myent
entry .input
text .mytxt
set list_of_entries [list .myent .mytxt .input]
foreach an_entry $list_of_entries {
    bind $an_entry <Key> [list text_binding_for_rtl %W %k]
}

[linuxpeter] - 2015-08-22 10:21:51

Thank you, Ahmad, for your great program! But it only works with unvowelled text, which is used in non-religious texts. While this text is displayed correctly in a Tk widget: مجانا لك This Bible quotation is not: كِتَابُ مِيلاَدِ يَسُوعَ الْمَسِيحِ ابْنِ دَاوُدَ ابْنِ إِبْراهِيمَ: Can you think of a way to make the prog skip the vowel signs when computing?