a g@sRdZddlmZmZmZmZGdddeeZeeeeefedddZdS) zf Trie data structure. Designed to be used for tokenizer's dictionary, but can be for other purposes. )IterableIteratorListUnionc@seZdZGdddZeedddZedddd Zeddd d Zee ed d dZ ee dddZ e edddZedddZdS)Triec@seZdZdZddZdS)z Trie.NodeendchildrencCsd|_i|_dS)NFrselfr =/usr/local/lib/python3.9/dist-packages/pythainlp/util/trie.py__init__szTrie.Node.__init__N)__name__ __module__ __qualname__ __slots__rr r r r Node srwordscCs,t||_t|_|D]}||qdSN)setrrrrootadd)r rwordr r r rs  z Trie.__init__N)rreturncCsT|}|j||j}|D]*}|j|}|sDt}||j|<|}qd|_dS)z Add a word to the trie. Spaces in front of and following the word will be removed. :param str text: a word TN) striprrrr getrrr)r rcurchchildr r r rs   zTrie.addcCs||jvrdS|j||j}g}|D]"}|j|}||||f|}q(d|_t|D]"\}}}|jsp|jrtq~|j|=qZdS)z} Remove a word from the trie. If the word is not found, do nothing. :param str text: a word NF)rremoverr appendrreversed)r rparentdatarr r r r r!-s    z Trie.remove)textrcCsTg}|j}t|D]<\}}|j|}|s.qP|jrJ||d|d|}q|S)z List all possible words from first sequence of characters in a word. :param str text: a word :return: a list of possible words :rtype: List[str] N)r enumerater rrr")r r&resrirnoder r r prefixesGs z Trie.prefixes)keyrcCs ||jvSrr)r r-r r r __contains__ZszTrie.__contains__)rccs|jEdHdSrrr r r r __iter__]sz Trie.__iter__cCs t|jSr)lenrr r r r __len__`sz Trie.__len__)rrrrrstrrrr!rr,boolr.rr/intr1r r r r r sr) dict_sourcercCstg}t|trbt|dkrbt|ddd$}|}t|}Wdq1sV0Yn&t|trt|tst|}ntd|S)a  Create a dictionary trie from a file or an iterable. :param str|Iterable[str]|pythainlp.util.Trie dict_source: a path to dictionary file or a list of words or a pythainlp.util.Trie object :return: a trie object :rtype: pythainlp.util.Trie rrutf8)encodingNziType of dict_source must be pythainlp.util.Trie, or Iterable[str], or non-empty str (path to source file)) r isinstancer2r0openread splitlinesr TypeError)r5ZtriefZ_vocabsr r r dict_trieds  (  r?N) __doc__typingrrrrr2rr?r r r r sX