diff --git a/addon/search_engine.py b/addon/search_engine.py index a64b20a..81734d0 100644 --- a/addon/search_engine.py +++ b/addon/search_engine.py @@ -52,15 +52,10 @@ def csv_to_list(csv): ] sql_joins_txt = "".join(joins) - -# we ignore these radicals for now because they don't have a keyword containing -# kanji counterpart in our database -ignore_radicals = ['亅','屮','禸','爻','尢','无','彑','黽','鬲','鬥','齊','龠','黹','黹','鬯'] - -# this table links radicals to their Heisig primitive counterpart to allow for -# even better search capability -# (some of them have identical visual look but differ only in Unicode) -unicode_conversion_table = { +# This conversion table links radicals to their Heisig primitive counterpart to allow +# for better search capability. Few of them have identical visual look but differ +# only in Unicode. +radical_conversion_table = { 'ノ' : '丿', '｜' : '丨', '⺅' : '亻', @@ -70,7 +65,7 @@ def csv_to_list(csv): } # When two characters reference each other as alternatives (for example 艹 -> 艸 and 艸 -> 艹 ) -# then we want to link to the character which is the primary primitive +# then we want to link to the character which is the primary primitive. primary_primitives = ['艹','扌','⻖','⻏','川','罒','冫','月'] class SearchEngine: @@ -99,8 +94,8 @@ def __init__(self, db_cursor): def radical_to_primitive(self,r): # First do unicode conversion because some radicals in the list might use slightly # # different (albeit visually indistinguishable) unicode character. - if r in unicode_conversion_table: - r = unicode_conversion_table[r] + if r in radical_conversion_table: + r = radical_conversion_table[r] # .. then reference the main primitive instead if this is an alternative primitive if r not in primary_primitives and r in self.primitive_alternative_cache: r = self.primitive_alternative_cache[r] @@ -131,7 +126,6 @@ def update_recursive_primitive_cache(self,character): if p in self.keyword_set_cache: kw_set = self.keyword_set_cache[p] for kw in kw_set: - #if kw not in rec_primitive_names_list: rec_primitive_names_list.append(kw) else: print("Note! Kanji %s references primitive %s without a keyword" % (character,p)) @@ -154,9 +148,6 @@ def init_cache(self): r = self.radical_to_primitive(r) if r in self.keyword_set_cache: radical_names_set.update(self.keyword_set_cache[r]) - else: - if r not in ignore_radicals: - print("Note! Kanji %s has unknown radical %s" % (c,r)) radical_names = ','.join(radical_names_set) self.radical_name_set_cache[c] = radical_names_set self.radical_name_cache[c] = radical_names @@ -209,7 +200,7 @@ def update_cache(self, character=None): if len(d['primitives'])>0: self.primitive_list_cache[c] = custom_list(d['primitives']) - # Radicals.. + # Radicals.. if len(d['radicals'])>0: self.radical_set_cache[c] = set(custom_list(d['radicals'])) @@ -238,9 +229,9 @@ def update_cache(self, character=None): st = d['usr_story'].lower() + d['koohi_stories'].lower() st += d['heisig_story'].lower() st += d['heisig_comment'].lower() - self.stories_cache[c] = st + # Frequency ranking points: Prioritize high frequency (and higher Kanken grade) kanjis. points = 0 if d['frequency_rank'] is not None and d['frequency_rank'] != '': fr_points = (4000 - d['frequency_rank'])/400 @@ -269,10 +260,9 @@ def update_cache(self, character=None): self.update_recursive_primitive_cache(c) - - def get_matching_characters(self, search_terms, pool, is_set, results, max_results): + def get_matching_characters(self, search_terms, pool, is_a_set, results, max_results): for search_term,required_count in search_terms.items(): - if is_set and required_count>1: + if is_a_set and required_count>1: # we want more than 1 occurence but this is a set -> not found return results @@ -293,9 +283,9 @@ def get_matching_characters(self, search_terms, pool, is_set, results, max_resul return results - def get_matching_characters_with_scoring(self, search_terms, pool, is_set, pool_priority, kanji_scores, kanji_matches): + def get_matching_characters_with_scoring(self, search_terms, pool, is_a_set, pool_priority, kanji_scores, kanji_matches): for search_term,required_count in search_terms.items(): - if is_set and required_count>1: + if is_a_set and required_count>1: # we want more than 1 occurence but this is a set -> not found return @@ -323,12 +313,13 @@ def get_matching_characters_with_scoring(self, search_terms, pool, is_set, pool_ def get_matching_characters_from_list_of_pools(self, search_terms, pool_list, max_results): if len(search_terms) == 1: - # In the case of only one search term its a simple exhaustive search until enough matches are found. - # Search goes through all search pools (keywords, primitive names, free text search) starting + # In the case of only one search term its a matter of simple exhaustive search + # until enough matches are found. Search crams through all search pools + # (keywords, primitive names, free text search) starting # from the most prioritized one results = [] - for pool_priority, pool, is_set in pool_list: - self.get_matching_characters(search_terms, pool, is_set, results, max_results) + for pool_priority, pool, is_a_set in pool_list: + self.get_matching_characters(search_terms, pool, is_a_set, results, max_results) if len(results)>=max_results: return results return results @@ -339,8 +330,8 @@ def get_matching_characters_from_list_of_pools(self, search_terms, pool_list, ma kanji_scores = dict() # score for each kanji kanji_matches = dict() # how many search terms were matched - for pool_priority, pool, is_set in pool_list: - self.get_matching_characters_with_scoring(search_terms, pool, is_set, pool_priority, kanji_scores, kanji_matches) + for pool_priority, pool, is_a_set in pool_list: + self.get_matching_characters_with_scoring(search_terms, pool, is_a_set, pool_priority, kanji_scores, kanji_matches) # remove those kanjis that didn't match all the search terms for kanji, matched_search_terms in kanji_matches.items(): @@ -389,7 +380,8 @@ def search(self, search_str, max_results=15): search_terms_dict[term] = 1 # A list of search pools, each having a distinct priority - priority_list = [ + priority_list = [ + # [ priority, pool, the_pool_is_a_set ] [30,self.keyword_set_cache,True], [26,self.keyword_cache,False], [20,self.rec_primitive_list_cache,False],