9eddlmZddlmZddlmZddlmZddl m Z ddl m Z  dd lZgd ZGd d e ZGdde ZGdde ZeZeZeZy #e$r dd lmZd ZYJwxYw)) annotations)SequenceMatcher)Any) find_ngrams)BaseSimilarity)TestFuncN)array)lcsseqlcsstrratcliff_obershelpLCSSeqLCSStrRatcliffObershelpcNeZdZdZ d d dZd dZd dZd dZd dZy) rzplongest common subsequence similarity https://en.wikipedia.org/wiki/Longest_common_subsequence_problem NcJ||_|xs |j|_||_yN)qval_ident test_funcexternal)selfrrrs H/usr/lib/python3/dist-packages/textdistance/algorithms/sequence_based.py__init__zLCSSeq.__init__"s"  "1dkk  c tr6tjt|dzt|dzft}n>t t|dzDcgc]}t ddgt|dzz }}t |D]_\}}t |D]L\}}||k(r|||dz||dz|dz<#t||dz||||dz||dz|dz<Nad} t|t|}}|dk7rl|dk7rg|||||dz |k(r|dz}nB||||||dz k(r|dz}n(||dz ||dz k(sJ||dz | z} |dz}|dz}|dk7r|dk7rg| Scc}w)z https://github.com/chrislit/abydos/blob/master/abydos/distance/_lcsseq.py http://www.dis.uniroma1.it/~bonifaci/algo/LCSSEQ.py http://rosettacode.org/wiki/Longest_common_subsequence#Dynamic_Programming_8 r)dtypeLr)numpyzeroslenintranger enumeratemax) rseq1seq2lengths_ichar1jchar2results r_dynamiczLCSSeq._dynamic,s kk3t9q=#d)a-"@LGBGD TU BVWQuS1#TQ"78WGW"$ VHAu%dO V5E>,3AJqMA,=GAEN1q5),/Aq0A71:aRSeCT,UGAEN1q5)  V V4y#d)11faqz!}Aq 11QA'!*QU"33QAE{d1q5k111a!ev-QQ1fa /Xs#E3crt|st|dS|j|Dcgc]}|d c}r"|dd}td|D}|||zSt|d}t |D]0\}}|d||ddfz||dzdz}t |||gt }2|Scc}w)Nrc3&K|] }|dd yw)Nr3).0ss r z$LCSSeq._recursive..Ts8af8srkey)alltypertupler&r'r#)r sequencesr7cmr,sss r _recursivezLCSSeq._recursiveOs9~%4 ! %' ' 4>>95aAbE5 6! R A8i88I#a' ' D1  i( -DAq2A!CR&+iA.??BT2YN,A -6s B4c~|sy|j|}t|dk(r|j|S|j|S)Nr r)_get_sequencesr#r1rBrr>s r__call__zLCSSeq.__call__\sH'D''3 y>Q  4==), ,"4??I. .rct||Srr#rEs r similarityzLCSSeq.similaritye4#$$r)rNT)rr$rr rboolreturnNone)r(strr)rNrLrNr>rNrLrN)rLr$) __name__ __module__ __qualname____doc__rr1rBrFrIr5rrrrsR" !!! !  !!F /%rrc0eZdZdZddZddZddZd dZy) rz(longest common substring similarity ct||}|jdt|dt|}||j|j|jzS)N)abr)_SequenceMatcherfind_longest_matchr#rVsize)rs1s2matchermatchs r _standartzLCSStr._standartmsJ"R2.**1c"gq#b'B%''577UZZ/00rct|t}t|}t|ddD]6}t||D]%}dj |}|D]}||vs |ccS8t |S)Nr9rr3r )minr#r%rjoinr<)rr>shortlengthnsubseqjoinedseqs r_customzLCSStr._customrsI3'Uvq"% "A%eQ/ "$"CS(""M  " "tE{}rct|syt|}|dk(ry|dk(r|dS|j|}|dk(r+tt t|dkr|j |S|j |S)Nr rrr)r;r#rDr'mapr_ri)rr>rds rrFzLCSStr.__call__s~9~Y Q; Q;Q< 'D''3 Q;3s3 23c9!4>>9- -t||Y''rct||SrrHrEs rrIzLCSStr.similarityrJrN)r[rNr\rNrLrNrOr>rNrLr$)rPrQrRrSr_rirFrIr5rrrris1  (%rrc(eZdZdZddZddZddZy)ra"Ratcliff-Obershelp similarity This follows the Ratcliff-Obershelp algorithm to derive a similarity measure: 1. Find the length of the longest common substring in sequences. 2. Recurse on the strings to the left & right of each this substring in sequences. The base case is a 0 length common substring, in which case, return 0. Otherwise, return the sum of the current longest common substring and the left & right recursed sums. 3. Multiply this length by 2 and divide by the sum of the lengths of sequences. https://en.wikipedia.org/wiki/Gestalt_Pattern_Matching https://github.com/Yomguithereal/talisman/blob/master/src/metrics/ratcliff-obershelp.js https://xlinux.nist.gov/dads/HTML/ratcliffObershelp.html cy)Nrr5rEs rmaximumzRatcliffObershelp.maximumsrc t|}t|}|dk(ry|Dcgc]}|d|j|}}|Dcgc]}||j||zd}}|j||z|j|zScc}wcc}w)Nr)rr#find_find)rr>rfrdr7beforeafters rrtzRatcliffObershelp._finds9%V Q;.78!OQVVF^$886?@166&>F*+,@@tzz6"V+jdjj%.@@@9@s BB c|j|}||St|}ttt|}|j|}||j |z|z Sr) quick_answerr#sumrlrDrt)rr>r0scountecounts rrFzRatcliffObershelp.__call__sf"""I.  MYSi()'D''3   I..77rNrn)r>rNrLfloat)rPrQrRrSrqrtrFr5rrrrs A8rr) __future__rdifflibrrXtypingrutilsrbaser _BaseSimilaritytypesr r! ImportErrorr __all__rrrr r rr5rrrs"8 3  J%_J%Z%%_%%P$8$8N  &(U EsA%% A54A5