ó F»]@c@sVdZddlZddlZd„Zd„Zd„ZejejZd„Z dS(sÏ Given a trie, find all occurrences of a word in the trie in a string. Like searching a string for a substring, except that the substring is any word in a trie. Functions: match Find longest key in a trie matching the beginning of the string. match_all Find all keys in a trie matching the beginning of the string. find Find keys in a trie matching anywhere in a string. find_words Find keys in a trie matching whole words in a string. iÿÿÿÿNcCs`d}xStt|ƒƒD]?}||d }|j|ƒs@Pn|j|ƒr|}qqW|S(s„match(string, trie) -> longest key or None Find the longest key in the trie that matches the beginning of the string. iN(tNonetrangetlent has_prefixthas_key(tstringttrietlongesttitsubstr((s/oak/stanford/groups/akundaje/marinovg/programs/biopython-1.50.tar.gz/biopython-1.50/build/lib.linux-x86_64-2.7/Bio/triefind.pytmatchs cCsgg}xZtt|ƒƒD]F}||d }|j|ƒs@Pn|j|ƒr|j|ƒqqW|S(s~match_all(string, trie) -> list of keys Find all the keys in the trie that matches the beginning of the string. i(RRRRtappend(RRtmatchesRR ((s/oak/stanford/groups/akundaje/marinovg/programs/biopython-1.50.tar.gz/biopython-1.50/build/lib.linux-x86_64-2.7/Bio/triefind.pyt match_all!scCswg}d}xd|t|ƒkrrt|||ƒ}x.|D]&}|j|||t|ƒfƒq;W|d7}qW|S(s‚find(string, trie) -> list of tuples (key, start, end) Find all the keys in the trie that match anywhere in the string. ii(RR R (RRtresultststarttkeystkey((s/oak/stanford/groups/akundaje/marinovg/programs/biopython-1.50.tar.gz/biopython-1.50/build/lib.linux-x86_64-2.7/Bio/triefind.pytfind1s $c Csítjdtjtƒƒ}g}d}x¾|t|ƒkrèt|||ƒ}xd|D]\}t|ƒ}||t|ƒks–|j|||ƒrW|j||||fƒqWqWW|j||ƒ}|dkrÙPn|j ƒ}q+W|S(sÌfind_words(string, trie) -> list of tuples (key, start, end) Find all the keys in the trie that match full words in the string. Word boundaries are defined as any punctuation or whitespace. s[%s]+iN( tretcompiletescapetDEFAULT_BOUNDARY_CHARSRR R R tsearchRtend( RRt _boundary_reRRRRtltm((s/oak/stanford/groups/akundaje/marinovg/programs/biopython-1.50.tar.gz/biopython-1.50/build/lib.linux-x86_64-2.7/Bio/triefind.pyt find_wordsCs  ! ( t__doc__RRR R Rt punctuationt whitespaceRR(((s/oak/stanford/groups/akundaje/marinovg/programs/biopython-1.50.tar.gz/biopython-1.50/build/lib.linux-x86_64-2.7/Bio/triefind.pyt s