#!/usr/bin/python # -*- coding = utf-8 -*- #************************************************************************ # $Id: verb_valid.py, v 0.7 2009/06/02 01:10:00 Taha Zerrouki $ # # ------------ # Description: # ------------ # Copyright (c) 2009, Arabtechies, Arabeyes Taha Zerrouki # # Elementary function to validate verbs # # ----------------- # Revision Details: (Updated by Revision Control System) # ----------------- # $Date: 2009/06/02 01:10:00 $ # $Author: Taha Zerrouki $ # $Revision: 0.7 $ # $Source: arabtechies.sourceforge.net # #***********************************************************************/ """ Basic routines to validate verbs ar_verb """ import re # import string # import sys # import os # import types # from arabic_const import * import libqutrub.verb_const as vconst #~ from verb_const import * # import ar_ctype import pyarabic.araby as araby from pyarabic.araby import FATHA, SHADDA, HAMZA, ALEF, \ NOON, ALEF_HAMZA_ABOVE, ALEF_HAMZA_BELOW, ALEF_MADDA, \ ALEF_MAKSURA, BEH, DAD, DAL, DAMMATAN, FATHATAN, FEH, GHAIN, HAH, \ HEH, JEEM, KAF, KASRATAN, KHAH, LAM, REH, SAD, SHEEN, TAH, TEH, \ TEH_MARBUTA, THAL, THEH, YEH, ZAH, ZAIN #used to VALID_INFINITIVE_VERB6_PATTERN = \ re.compile(u"^است...|ا..ن..|ا..و..|ا..ا.ّ|ا....ّ|ا.ّ.ّ.|ا.ّا..$", re.UNICODE) VALID_INFINITIVE_VERB4_PATTERN = re.compile(\ u"^([%s%s][^%s]{2}.|[^%s%s]%s[^%s%s].|[^%s%s]{2}%s[^%s]|[^%s%s]{4})$"\ %(ALEF_HAMZA_ABOVE, HAMZA, SHADDA, ALEF, SHADDA, ALEF, ALEF, SHADDA, ALEF, SHADDA, SHADDA, SHADDA, ALEF, SHADDA), re.UNICODE) VALID_INFINITIVE_VERB5_PATTERN = re.compile( u"|".join([ u"^ا...ّ$", # حالة اتخذ أو اذّكر أو اطّلع u"^%s[%s%s%s]%s..$"%(ALEF, TEH, THAL, TAH, SHADDA), # حالة اتخذ أو اذّكر أو اطّلع u"^ا[تذط]ّ[^اّ][^اّ]$", # انفعل u"^ان...$", #افتعل u"^(ازد|اصط|اضط)..$" u"^ا[^صضطظد]ت..$", u"^ا...ّ$", # حالة اتخذ أو اذّكر أو اطّلع u"^ا.ّ..$", u"^ا...ى$", ]) , re.UNICODE) ##################################### #{validation functions ##################################### def is_valid_infinitive_verb(word, vocalized = True): """ Determine if the given word is a valid infinitive form of an arabic verb. A word is not valid infinitive if - lenght < 3 letters. - starts with : ALEF_MAKSURA, WAW_HAMZA, YEH_HAMZA, HARAKAT - contains TEH_MARBUTA, Tanwin - contains non arabic letters. - contains ALEF_MAKSURA not in the end. - contains double haraka : a warning @param word: given word. @type word: unicode. @param is_vocalized: if the given word is vocalized. @type is_vocalized:Boolean, default(True). @return: True if the word is a valid infinitive form of verb. @rtype: Boolean. """ # test if the word is an arabic valid word, if not araby.is_arabicword(word): return False if vocalized : word_nm = araby.strip_harakat(word) else: word_nm = word # the alef_madda is considered as 2 letters word_nm = word_nm.replace(ALEF_MADDA, HAMZA+ALEF) length = len(word_nm) # lenght with shadda must be between 3 and 6 if length < 3 or length >= 7: return False # a 3 length verb can't start by Alef or Shadda, #and the second letter can't be shadda elif length == 3 and (word_nm[0] == ALEF or word_nm[0] == SHADDA \ or word_nm[1] == SHADDA): return False # a 5 length verb must start by ALEF or TEH elif length == 5 and word_nm[0] not in (TEH, ALEF): return False # a 6 length verb must start by ALEF elif length == 6 and word_nm[0] != ALEF: return False # contains some invalide letters in verb elif re.search(u"[%s%s%s%s%s]"%(ALEF_HAMZA_BELOW, TEH_MARBUTA, DAMMATAN, KASRATAN, FATHATAN), word): return False # contains some SHADDA sequence letters in verb # Like shadda shadda, shadda on alef, start # by shadda, shadda on alef_ maksura, # ALEF folowed by (ALEF, ALEF_MAKSURA) # ALEF Folowed by a letter and ALEF # end with ALEF folowed by (YEH, ALEF_MAKSURA) # first letter is alef and ALLw alef and two letters aand shadda elif re.search(u"([%s%s%s]%s|^%s|^%s..%s|^.%s|%s.%s|%s%s|%s[%s%s]$)"%( ALEF, ALEF_MAKSURA, SHADDA, SHADDA, SHADDA, ALEF, SHADDA, SHADDA, ALEF, ALEF, ALEF, ALEF, ALEF, ALEF_MAKSURA, YEH), word_nm): return False # Invalid root form some letters : #~ # initial YEH folowed by #~ ((THEH, JEEM, HAH, KHAH, THAL, ZAIN, SHEEN, SAD, DAD, #~ TAH, ZAH, GHAIN, KAF, HEH, YEH)) elif re.search(u"^%s[%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s]"%( YEH, THEH, JEEM, HAH, KHAH, THAL, ZAIN, SHEEN, SAD, DAD, TAH, ZAH, GHAIN, KAF, HEH, YEH), word_nm): return False # TEH After (DAL, THAL, TAH, ZAH, DAD) elif re.search(u"[%s%s%s%s%s]%s"%(DAL, THAL, DAD, TAH, ZAH, TEH), word_nm): return False # Contains invalid root sequence in arabic, near in phonetic # like BEH and FEH, LAM And REH elif re.search(u"%s%s|%s%s|%s%s|%s%s|%s%s|%s%s|%s%s"%( LAM, REH, REH, LAM, FEH, BEH, BEH, FEH, NOON, LAM, HEH, HAH, HAH, HEH), word_nm): return False # in non 5 letters verbs :initial TEH followed by # (THEH, DAL, THAL, ZAIN, SHEEN, SAD, DAD, TAH, ZAH) elif length != 5 and word_nm.startswith(TEH) and word_nm[1] in ( TEH, THEH, DAL, THAL, ZAIN, SHEEN, SAD, DAD, TAH, ZAH): return False # if word start by the same letter doubled elif word_nm[0] == word_nm[1] and word[0] != TEH: return False #verify the wazn of the verb elif length == 3: if re.match("^[^%s][^%s].$"%(ALEF, SHADDA), word_nm): return True # الأوزان المقبولة هي فعل، فعّ، # الأوزان غير المقبولة # اعل، فّل else: return False elif length == 4: #1- أفعل، 2- فاعل، 3 فعّل 4 فعلل if re.match(\ "^([%s%s][^%s]{2}.|[^%s%s]%s[^%s%s].|[^%s%s]{2}%s[^%s]|[^%s%s]{4})$"\ %(ALEF_HAMZA_ABOVE, HAMZA, SHADDA, ALEF, SHADDA, ALEF, ALEF, SHADDA, ALEF, SHADDA, SHADDA, SHADDA, ALEF, SHADDA), word_nm): return True # الأوزان المقبولة هي فعل، فعّ، # الأوزان غير المقبولة # افعل: يجب تثبيت همزة القطع #فّعل، فعلّ: الشدة لها موضع خاص # فعال، فعلا: للألف موضع خاص else: return False elif length == 5: if word_nm.startswith(ALEF): if re.match(u"^ا...ّ$", word_nm): return True # حالة اتخذ أو اذّكر أو اطّلع if re.match(u"^%s[%s%s%s]%s..$"%(ALEF, TEH, THAL, TAH, SHADDA), \ word_nm): return True # انفعل elif re.match(u"^ان...$", word_nm): return True #افتعل elif re.match(u"^(ازد|اصط|اضط)..$", word_nm): return True elif re.match(u"^ا[^صضطظد]ت..$", word_nm): return True elif re.match(u"^ا...ّ$", word_nm): return True # حالة اتخذ أو اذّكر أو اطّلع elif re.match(u"^ا.ّ..$", word_nm): return True elif re.match(u"^ا...ى$", word_nm): return True else: return False elif word_nm.startswith(TEH): return True else: return False # الأوزان المقبولة هي فعل، فعّ، # الأوزان غير المقبولة #للشدة موضع خاص: تفعّل، افتعّ # للألف مواضع خاصة، elif length == 6: if not (word_nm.startswith(ALEF) or word_nm.startswith(TEH)): return False if VALID_INFINITIVE_VERB6_PATTERN.match(word_nm): return True # الأوزان المقبولة هي فعل، فعّ، # الأوزان غير المقبولة #للشدة موضع خاص: تفعّل، افتعّ # للألف مواضع خاصة، else: return False return True def suggest_verb(verb): """ Generate a list of valid infinitive verb for an invalid infinitive form. @param verb: given verb, of invalid infinitive form. @type verb: unicode. @return: a list of suggested infinitive verb forms @rtype: list of unicode. """ # the verb is invalid list_suggest = [] # first strip harakat, shadda is not striped verb = araby.strip_harakat(verb) # second strip all inacceptable letters in an infinivive form verb = re.sub(u"[%s%s%s%s]"%( TEH_MARBUTA, DAMMATAN, KASRATAN, FATHATAN), \ '', verb) # test the resulted verb if it's valid, if ok, # add it to the suggestion list. if is_valid_infinitive_verb(verb): list_suggest.append(verb) return list_suggest # if the verb starts by ALEF_HAMZA_BELOW like إستعمل, #replace if by an ALEF, because it's a common error. # if the result is valid add it to the suggestions list elif verb.startswith(ALEF_HAMZA_BELOW): verb = re.sub(ALEF_HAMZA_BELOW, ALEF, verb) if is_valid_infinitive_verb(verb): list_suggest.append(verb) return list_suggest # if the verb starts by ALEF like اضرب, #replace if by an ALEF_HAMZA_ABOVE, because it's a common error. # if the result is valid add it to the suggestions list elif verb.startswith(ALEF): verb_one = re.sub(ALEF, ALEF_HAMZA_ABOVE+FATHA, verb, 1) if is_valid_infinitive_verb(verb_one): list_suggest.append(verb_one) return list_suggest # if the verb is 2 letters length, # suggest to add the third letter as : # Shadda, Alef, Alef Maksura, Yeh at the end # if the result is valid add it to the suggestions list elif len(verb) == 2: verb = re.sub(ALEF, ALEF_HAMZA_ABOVE, verb, 1) #suggest to add the third letter as : Shadda at the end verb_one = verb+SHADDA if is_valid_infinitive_verb(verb_one): list_suggest.append(verb_one) #suggest to add the third letter as : Alef Maksura verb_one = verb+ALEF_MAKSURA if is_valid_infinitive_verb(verb_one): list_suggest.append(verb_one) #suggest to add the third letter as :Alef at the end verb_one = verb+ALEF if is_valid_infinitive_verb(verb_one): list_suggest.append(verb_one) #suggest to add the third letter as :Alef in middle verb_one = verb[0]+ALEF+verb[1] if is_valid_infinitive_verb(verb_one): list_suggest.append(verb_one) return list_suggest elif len(verb) >= 6: # if the verb is more than 6 letters length, #suggest to replace the over added letters by Alef # if the result is valid add it to the suggestions list for i in range(len(verb)-6): verb_one = ALEF+verb[i:i+5] if is_valid_infinitive_verb(verb_one): list_suggest.append(verb_one) elif len(verb) == 5: # if the verb is 5 letters length, suggest # if the result is valid add it to the suggestions list # ToDo: review this part for i in range(len(verb)-5): verb_one = ALEF+verb[i:i+4] if is_valid_infinitive_verb(verb_one): list_suggest.append(verb_one) elif len(verb) == 4: # if the verb is 5 letters length, #suggest to replace the over added letters by Alef # if the result is valid add it to the suggestions list # فعال = > فاعل #فّعل = > فعّل if verb[2] == ALEF or verb[1] == SHADDA: verb_one = verb[0]+verb[2]+verb[1]+verb[3] if is_valid_infinitive_verb(verb_one): list_suggest.append(verb_one) if verb.endswith(SHADDA): # if the verb is 4 letters length, #suggest to correct the alef and shadda position # if the result is valid add it to the suggestions list #فعلّ = > فعّل verb_one = verb[0]+verb[1]+verb[3]+verb[2] if is_valid_infinitive_verb(verb_one): list_suggest.append(verb_one) return list_suggest else: # else sugest to conjugate another verb list_suggest.append(u"كتب") return list_suggest return list_suggest ##################################### #{verb pretreatment functions ##################################### def normalize_alef_madda(word): """ Convert Alef madda into two letters. @param word: given word. @type word: unicode. @return: converted word. @rtype: unicode. """ if word.startswith(ALEF_MADDA): word_nm = araby.strip_harakat(word) if len(word_nm) == 2: return word_nm.replace(ALEF_MADDA, HAMZA+ALEF) elif len(word_nm) == 3: if vconst.ALEF_MADDA_VERB_TABLE.has_key(word_nm): #return the first one only #mylist = ALEF_MADDA_VERB_TABLE[word_nm] return vconst.ALEF_MADDA_VERB_TABLE[word_nm][0] else: return word_nm.replace(ALEF_MADDA, HAMZA+ALEF) else: return word_nm.replace(ALEF_MADDA, HAMZA+ALEF) else: return word_nm