Function include_in_dictionary

DeepPCB/evaluation/script.py:182–223

Function used in Word Spotting that finds if the Ground Truth transcription meets the rules to enter into the dictionary. If not, the transcription will be cared as don't care

(transcription)

                    
    
    def include_in_dictionary(transcription):
        """
        Function used in Word Spotting that finds if the Ground Truth transcription meets the rules to enter into the dictionary. If not, the transcription will be cared as don't care
        """        
        #special case 's at final
        if transcription[len(transcription)-2:]=="'s" or transcription[len(transcription)-2:]=="'S":
            transcription = transcription[0:len(transcription)-2]
        
        #hypens at init or final of the word
        transcription = transcription.strip('-');
        
        specialCharacters = unicode("'!?.:,*\"()·[]/","utf-8");
        for character in specialCharacters:
            transcription = transcription.replace(character,' ')
        
        transcription = transcription.strip()
        
        if len(transcription) != len(transcription.replace(" ","")) :
            return False;
        
        if len(transcription) < evaluationParams['MIN_LENGTH_CARE_WORD']:
            return False;
        
        notAllowed = unicode("×÷·","utf-8");
        
        range1 = [ ord(u'a'), ord(u'z') ]
        range2 = [ ord(u'A'), ord(u'Z') ]
        range3 = [ ord(u'À'), ord(u'ƿ') ]
        range4 = [ ord(u'DŽ'), ord(u'ɿ') ]
        range5 = [ ord(u'Ά'), ord(u'Ͽ') ]
        range6 = [ ord(u'-'), ord(u'-') ]
        
        for char in transcription :
            charCode = ord(char)
            if(notAllowed.find(char) != -1):
                return False
            
            valid = ( charCode>=range1[0] and charCode<=range1[1] ) or ( charCode>=range2[0] and charCode<=range2[1] ) or ( charCode>=range3[0] and charCode<=range3[1] ) or ( charCode>=range4[0] and charCode<=range4[1] ) or ( charCode>=range5[0] and charCode<=range5[1] ) or ( charCode>=range6[0] and charCode<=range6[1] )
            if valid == False:
                return False
        
        return True
    
    def include_in_dictionary_transcription(transcription):
        """