-iysdZddlZddlZddlZejdddlZddlmZ  ddl Z ddl Z n#ddl m Z ddlm Z YnxYwdZdZGdde j ZdS)z Arabic noun stemmer N.c$fd|DS)a@ Verify possible affixes in the resulted segments according to the given affixes list. @param word: the input word. @type word: unicode. @param list_seg: list of word segments indexes (numbers). @type list_seg: list of pairs. @return: list of acceped segments. @rtype: list of pairs. c~g|]9}dd|d|ddgv7|:S-Nrjoin.0s affix_listwords S/var/www/html/speakWrite/venv/lib/python3.11/site-packages/alyahmor/noun_affixer.py z verify_affix..+sW    88T%1Q4%[$qtuu+. / /: = = = = =rlist_segrs` `r verify_affixr s4        rc`|}d|vr |drdS|ddvrd|vrdSd|vrdSd|vrdS|d vr |d sdS|d vr |d sdSd |vr |dsdS|d r |dsdS|tjkr |dsdS|tjkr|ddkrdSdS)aF Test if the given word from dictionary is compabilbe with affixes tags. @param noun_tuple: the input word attributes given from dictionary. @type noun_tuple: dict. @param affix_tags: a list of tags given by affixes. @type affix_tags:list. @param proclitic_nm: first level prefix vocalized. @type proclitic_nm: unicode. @param enclitic_nm: first level suffix vocalized. @type enclitic_nm: unicode. @param suffix_nm: first level suffix vocalized. @type suffix_nm: unicode. @return: if the tags are compatible. @rtype: Boolean. تنوين mamnou3_sarfFnumber)جمعuجمع تكسيرuجمع مؤنث سالمuجمع مذكر سالمuمثنى)uهمuهنuكماuكمuهما hm_suffix)uهuها ha_suffixالk_prefix kal_prefixw_suffixrT)endswitharWAW TEH_MARBUTA) noun_tuple affix_tags proclitic_nm enclitic_nm suffix_nm proclitics r validate_tagsr,0s,$I ""z.'A"u(@@@ &* 4 45 &* 4 45 * $ $5"""*4[*A"u&&&z+/F&ui :(>u'"":l+CuBF:j#9uBN""z(';y'H'Hu 4rceZdZdZedZddZedZedZdZ edZ ed Z ed Z dd Z d ZdZgfdZddZdZdS) noun_affixercptj|tj|_g|_tj|_tj|_ tj |_ tj |_ tj|_g|_tj|_tj|_dS)N) basic_affixer__init__SNCCOMP_PREFIX_LIST procleticsprefixesCONJ_SUFFIX_LISTsuffixesCOMP_SUFFIX_LIST encliticsNOMINAL_CONJUGATION_AFFIXaffixesCOMP_NOUN_AFFIXEScliticsCOMP_PREFIX_LIST_TAGSprocletics_tags prefixes_tagsCONJ_SUFFIX_LIST_TAGS suffixes_tagsCOMP_SUFFIX_LIST_TAGSenclitics_tags)selfs rr1znoun_affixer.__init__ls#,,T333. , -4 ,  #8 6!7rct|g}|tjtjztjtjztjtjtjztjzfvr$|tjz}|||r4|tjtjztjtjzfvr$|tjz}||| tjr,|ddtj z}||| tj r,|ddtj z}|||}|S)u Generate the Noun stem variants according to the affixes. For example مدرستي = >مدرست+ي = > مدرسة +ي. Return a list of possible cases. @param stem: the input stem. @type stem: unicode. @param suffix_nm: suffix (no mark). @type suffix_nm: unicode. @return: list of stem variants. @rtype: list of unicode. N) setr#ALEFTEHYEHr%addNOONr$r" ALEF_MAKSURAHAMZA YEH_HAMZA)stemr*possible_noun_list possible_nounvalidated_lists rget_stem_variantsznoun_affixer.get_stem_variantssN"! "    26)26BN+BBF"')BF24 4 4 2>1M  " "= 1 1 1 2I"&27*:BFRW>II (f!5!5 (BCCyRVRW---  (K  D.v6v> > >#%#6y#A#A #, ...rc|}|tjtjzr|tjr{|tjtjztj}|tjtjztjtjz}|tjr|tjtjzs|tjr%tj }tj tj z}|tjtj ztj zsr|tjtj ztjzs9|tjtj ztjzrd}d}||fS)u Get the enclitix variant to be joined to the word. For example: word = كتاب, suffix = كسرة, enclitic = هم. The enclitic has a second form هِم. @param enclitic_voc: first level suffix vocalized. @type enclitic_voc: unicode. @param suffix_voc: second level suffix vocalized. @type suffix_voc: unicode. @return: variant of enclitic (vocalized enclitic and vocalized enclitic without I'rab short mark). @rtype: (unicode, unicode) rW) startswithr#HEHDAMMAr"KASRAreplacerKSUKUNSHADDAFATHA) enclitic_voc suffix_vocencl_vo_no_inflect_marks rget_enclitic_variantz!noun_affixer.get_enclitic_variants#/  " "26BH#4 5 5 C*:M:M;; C&2&:&:26BH;L;=6'C'C #'//0A020ACCL  " "26 * * # ""26"(?33 3z7J7J267R7R 3+-9'!y"(2 ##BF29$4bh$>?? #""26")#3RX#=>> #""26")#3RX#=>> #,/'" 444rcJ|}tj|}tj|}||z}tj|}|tjr|tjtjzr|tjtjtjztjtjztjzfvr|ddtjz}n.|tjtjzkrtj|dd}n|dkr|ddtjz}n|tjtjztjtjztjtjtjztjzfvrtj|dd}nR|dkr|ddtjz}n3|tjr>|dkr|ddtjz}n|dkr|ddtjz}n|tj tjzr|d|vr&d|vr"|s |stj|dd}n|tjtj ztj tj zfvrtj|dd}n.|tj r|s|r | tjr|ddtjz}n| tj r|ddtjz}n|tjtj zs9|tjtjztj zr6| tjr|ddtjz}|S)u Get the word variant to be joined to the suffix. For example: word = مدرسة, suffix = ي. The word is converted to مدرست. @param word: word found in dictionary. @type word: unicode. @param proclitic: proclitic ( first level). @type proclitic: unicode. @param suffix: suffix ( first level). @type suffix: unicode. @param enclitic: enclitic( second level). @type enclitic: unicode. @return: variant of word. @rtype: unicode. NrGrWruلل)r#strip_tashkeelr^r"r%rIrKrJrNrirMr$rOrfrh WAW_HAMZArPrkFATHATAN)rr+r_r` word_stemr*r)long_suffix_nms rget_word_variantznoun_affixer.get_word_variants  %f-- '11 "[0' 22    bn - -3 :!!"'BN":;; 4".)@!#"'!1BF!:!<<< )#2# 7II"'BF"222 " 3IcrcN C CII#s** )#2# 7Irw/".1H"&!v/"&8:::/ #2#??  3&&%crcNRV3    0 0 :C%crcNRV3 ##%crcNRW4   26 1 2 2 :i''G9,D,D[,Daj,D/ #2#?? rv/"'1ABBB/ #2#??    ) ) :y :K :  ** :%crcNR\9 ""28,, :%crcNR\9 $$RVbh%677 :''(9BH(DEE :((55 :&crcNR\9 rchg}|}tj|ddr |dd}|tjtj}|tjtjtjz}|tjtjtjz}tjdtjztj|}tjdtjzd|}| tj tj zrdnd}| ||||}| ||||\}}|j|id|D]} |j|id|D]} || |\} } d |j|d vrstj|d rYd|d tj|d dg}| tjr | dd} d| ||| g} d| ||| g} d | ||| g}tj|}t.jD]1\}}| ||} | ||} 2|| | |f|S)a Join the noun and its affixes, and get the vocalized form @param noun: noun found in dictionary. @type noun: unicode. @param proclitic: first level prefix. @type proclitic: unicode. @param suffix: second level suffix. @type suffix: unicode. @param enclitic: first level suffix. @type enclitic: unicode. @return: vocalized word. @rtype: unicode. rGNz(%s)+z^(%s)+rWTF vocalized تعريفrXrrr)r#r\rjrvrmrIrNrZr[r"rirKryrdr?getrDrqis_sunr rlrkrtr2AJUST_VOCAL_PATTERNSappend)rEnounr+r_r`word_tuple_listrorarc proclitic_vocrnencl_voc_non_inflectword_non_irab_markword_vocalized segmentedpatrnrepls rvocalizeznoun_affixer.vocalizeas   <RSS " " 9D||BK22 ||BGRX%788||BORX-GHHvh)28T::vi"(*B55--26(9::E$$$$T9fhGG ,0+C+C *h,1,1( ("155iCCGG U^__ T TM $ 3 7 7 D D H HV^ _ _ * * 595N5N &6*6*2 22  4Y ? GGGBIVZ[\V]L^L^Gxxa")T!""X >?? ))"(337$1#2#$6M "$&:> @ @ @rczg}|r|tjr+||ddtjznl|tjr*||ddtjzn"||tjz|S)z generate stem varaintes NrG)r"r#rIrrNrJr%rK)rQr) list_stems rget_input_stem_variantsz$noun_affixer.get_input_stem_variantss  ,}}RW%% =  crcR_!<====rv&& =  crcR^!;<<<   TBF] + + +rc$fd|DS)ad Verify possible affixes in the resulted segments according to the given affixes list. @param word: the input word. @type word: unicode. @param list_seg: list of word segments indexes (numbers). @type list_seg: list of pairs. @return: list of acceped segments. @rtype: list of pairs. c~g|]9}dd|d|ddgv7|:Srr r s rrz-noun_affixer.verify_affix..sW   xxeqted1Q455k233zAA AAArrrs` `rrznoun_affixer.verify_affixs4         rrWc 8g}tj|}|||rm||||rV|||||}|r.s#B#B#BDGG#B#B#Br)r#rtis_valid_cliticscheck_clitic_affixrrget_tags)rErprocprefsuffenc newword_list word_tuples rget_formznoun_affixer.get_forms  &&  s + + Q&&tS$77 Q#}}T4sCC Q#B#B\#B#B#BL&2QQ "))$--tT3*O*OPPPPrcg}tj|j|j|jD]J}|d}|d}|d}|||d||}|r||K|S)z generate all possible affixesrrrrW) itertoolsproductr4r7r9rextend)rEr noun_formselementrrrrs rgenerate_formsznoun_affixer.generate_formss  ( t~'' 0 0G1:D1:D!*C==tRcBBL 0!!,///rcg}|j|idd}|||j|idd}|||j|idd}||d|D}d|S)z+ Get affixes tags rXrcg|]}||Srr)r ts rrz)noun_affixer.get_tags.."s++++1+++r:)r?r}rrDrBr ) rEr procleticr_r`taglistproclitic_tags enclitic_tags suffix_tagss rrznoun_affixer.get_tagss -11)R@@DDVBOO~&&&+//"==AA&LL }%%%(,,VR88<.?s'IIIR.qt44IIIrcg|] }|d Srrrs rrz4noun_affixer.generate_affix_list..As666aQqT666rc<g|]}|dS)r)rj)r rrs rrz4noun_affixer.generate_affix_list..Es'CCC4,,CCCr)rrrH)rEr{r list_affixesrs @rgenerate_affix_listz noun_affixer.generate_affix_list8s((..  7IIjIIILL66:666LC --.. CCCClCCC rch|j|idd}|j|idd}|j|idd}||j|iddz }d|vrd|vrdSd|vrd|vrdSd|vrd|vrdSd|vrd |vrdSd |vrd |vrdS|t jr!|t jrdSd S) a Verify if proaffixes (sytaxic affixes) are compatable with affixes ( conjugation) @param proclitic_nm: first level prefix. @type proclitic_nm: unicode. @param enclitic: first level suffix. @type enclitic: unicode. @param suffix: second level suffix. @type suffix: unicode. @return: compatible. @rtype: True/False. rXrcasesr|rFu إضافةuمضافu لايضافuجرu مجرورT) r?r}rDrBrfr#rKr"rh)rEr(r`r_rrrs rrznoun_affixer.check_clitic_affixHskN-11,CCGGrRR+//"==AA&LL (,,VR88<rs /%%%%%/222222........ 777vttttt=.ttttts =A