Python query.Term方法代码示例

# 需要导入模块: from whoosh import query [as 别名]
# 或者: from whoosh.query import Term [as 别名]
def search(self, w: str) -> SearchResult:
        if not self.ix.up_to_date():
            self.initialize_trie() # if the index is not up to date, someone has added cards, so we reinitialize the trie

        normalized = list(WhooshConstants.normalized_analyzer(w))[0].text

        # If we get matches by prefix, we return that
        exact, prefix_whole_word, other_prefixed = self.find_matches_by_prefix(normalized)
        if exact or len(prefix_whole_word) > 0 or len(other_prefixed) > 0:
            return SearchResult(exact, prefix_whole_word, other_prefixed, [])

        # We try fuzzy and stemmed queries
        query_normalized = fuzzy_term(normalized, self.DIST, 'name_normalized')
        query_stemmed = And([Term('name_stemmed', q.text) for q in WhooshConstants.stem_analyzer(w)])
        query_tokenized = And([fuzzy_term(q.text, self.DIST, 'name_tokenized') for q in WhooshConstants.tokenized_analyzer(w)])
        if len(query_tokenized) == 0: # This can be empty because some unicode chars are ignored. See #4988
            query = Or([query_normalized, query_stemmed])
        else:
            query = Or([query_normalized, query_tokenized, query_stemmed])

        with self.ix.searcher() as searcher:
            fuzzy = [(r['canonical_name'], r.score) for r in searcher.search(query, limit=40)]
        return SearchResult(exact, prefix_whole_word, other_prefixed, fuzzy) 

# 需要导入模块: from whoosh import query [as 别名]
# 或者: from whoosh.query import Term [as 别名]
def fuzzy_term(q: str, dist: int, field: str) -> Term:
    if len(q) <= 3:
        return Term(field, q)
    return FuzzyTerm(field, q, maxdist=dist, prefixlength=1) 

# 需要导入模块: from whoosh import query [as 别名]
# 或者: from whoosh.query import Term [as 别名]
def get_all_num(self, keyword, catid=''):
        queryit = self.parser.parse(keyword)
        if catid == '':
            pass
        else:
            queryit = And([Term("catid", catid), queryit])

        results = self.whbase.searcher().search(queryit)
        return len(results)
        # return len(self.whbase.searcher().search(queryit).docs()) 

# 需要导入模块: from whoosh import query [as 别名]
# 或者: from whoosh.query import Term [as 别名]
def search_pager(self, keyword, catid='', page_index=1, doc_per_page=10):

        queryit = self.parser.parse(keyword)
        if catid == '':
            pass
        else:
            queryit = And([Term("catid", catid), queryit])
        try:
            queryres = self.whbase.searcher().search(queryit, limit=page_index * doc_per_page)
            return queryres[(page_index - 1) * doc_per_page: page_index * doc_per_page]
        finally:
            pass 

# 需要导入模块: from whoosh import query [as 别名]
# 或者: from whoosh.query import Term [as 别名]
def search_addresses(searcher, query):
    restrict_q = Term("tag", "drafts") | Term("tag", "trash")
    results = []
    for field in ['to', 'cc', 'bcc', 'sender']:
        query_parser = QueryParser(field, searcher.schema)
        results.append(
            searcher.search(
                query_parser.parse("*%s* OR *%s*" % (query.title(), query)),
                limit=None,
                mask=restrict_q,
                groupedby=sorting.FieldFacet(
                    field,
                    allow_overlap=True),
                terms=True).matched_terms())
    return [address[1] for address in flatten(results)] 

# 需要导入模块: from whoosh import query [as 别名]
# 或者: from whoosh.query import Term [as 别名]
def cal_sim(train_data_path, test_data_path, dst_result_path=None, save_n_best_search=1):
    schema = Schema(context=TEXT(stored=True), response=STORED, post=TEXT(stored=True))
    index_i = re.findall('\d', train_data_path)[0]

    index_path = "../tmp/ix_index/" + index_i
    if not os.path.exists(index_path):
        os.makedirs(index_path)

    ix = create_in(index_path, schema)
    writer = ix.writer()

    def get_cpr(line):
        lines = line.lower().strip().split('\t')
        context = ''
        post = lines[0]
        response = lines[1]
        return context.strip().decode('utf-8'), response.decode('utf-8'), post.decode('utf-8')

    def load_train_data(file_name, writer):
        f = open(file_name)
        for line in f:
            context, response, post = get_cpr(line)
            if context != '':
                writer.add_document(context=context, response=response, post=post)
            else:
                writer.add_document(response=response, post=post)
        writer.commit()

    def get_query(line, ix):
        lines = line.strip().split('\t')
        post = lines[0].decode('utf-8')
        q2 = QueryParser("post", ix.schema).parse(post)
        terms = list(q2.all_terms())
        query = Or([Term(*x) for x in terms])
        return query

    load_train_data(train_data_path, writer)

    f = open(test_data_path, 'r')
    fw_search = open(dst_result_path, 'w')
    with ix.searcher(weighting=scoring.TF_IDF()) as searcher:
        c = searcher.collector(limit=10)
        tlc = TimeLimitCollector(c, timelimit=10.0)
        for line in f:
            try:
                query = get_query(line, ix)
                searcher.search_with_collector(query, tlc)
                results = tlc.results()
                for i in range(min(len(results), save_n_best_search)):
                    fw_search.write(
                        line.strip() + '\t' + str(results[i]["post"]) + '\t' + str(results[i]["response"]) + '\n')
            except Exception as e:
                print('TimeLimit, ignore it!')
                print(line)
    fw_search.close()