Небольшие правки

aore/fias/fiasfactory.py

@@ -1,5 +1,6 @@
 # -*- coding: utf-8 -*-
 from aore.fias.search import SphinxSearch
+import logging
 
 
 class FiasFactory:
@@ -7,11 +8,11 @@ class FiasFactory:
         self.searcher = SphinxSearch()
 
     # text - строка поиска
-    # strong - строгий поиск или "мягкий" (с допущением ошибок, опечаток)
+    # strong - строгий поиск (True) или "мягкий" (False) (с допущением ошибок, опечаток)
     # out_format - "full" or "simple" - полный (подробно для каждого подпункта) или простой (только строка и AOID)
     def find(self, text, strong=False, out_format="simple"):
         try:
             results = self.searcher.find(text, strong)
-
+            print results
         except:
             return []

aore/fias/search.py

@@ -5,6 +5,7 @@ import re
 import Levenshtein
 import psycopg2
 import sphinxapi
+import logging
 
 from aore.config import db as dbparams, sphinx_index_sugg, sphinx_index_addjobj
 from aore.dbutils.dbimpl import DBImpl
@@ -23,14 +24,14 @@ class SphinxSearch:
         self.db = DBImpl(psycopg2, dbparams)
 
         self.client_sugg = sphinxapi.SphinxClient()
-        self.client_sugg.SetServer("localhost", 9312)
+        self.client_sugg.SetServer("127.0.0.1", 9312)
         self.client_sugg.SetLimits(0, 10)
-        self.client_sugg.SetConnectTimeout(3.0)
+        self.client_sugg.SetConnectTimeout(7.0)
 
         self.client_show = sphinxapi.SphinxClient()
-        self.client_show.SetServer("localhost", 9312)
+        self.client_show.SetServer("127.0.0.1", 9312)
         self.client_show.SetLimits(0, 10)
-        self.client_show.SetConnectTimeout(3.0)
+        self.client_show.SetConnectTimeout(7.0)
 
     def __configure(self, index_name, wlen=None):
         if index_name == "idx_fias_sugg":
@@ -42,7 +43,7 @@ class SphinxSearch:
                 self.client_sugg.SetSortMode(sphinxapi.SPH_SORT_EXTENDED, "krank DESC")
         else:
             self.client_show.SetMatchMode(sphinxapi.SPH_MATCH_EXTENDED2)
-            self.client_show.SetRankingMode(sphinxapi.SPH_RANK_BM25)
+            #self.client_show.SetRankingMode(sphinxapi.SPH_RANK_BM25)
 
     def __get_suggest(self, word, rating_limit, count):
         word_len = str(len(word) / 2)
@@ -101,6 +102,7 @@ class SphinxSearch:
             if word != '':
                 we = WordEntry(self.db, word)
                 self.__add_word_variations(we, strong)
+
                 if we.get_variations() == "()":
                     raise BaseException("Cannot process sentence.")
                 yield we
@@ -111,7 +113,13 @@ class SphinxSearch:
         sentence = "{}".format(" MAYBE ".join(x.get_variations() for x in word_entries))
 
         self.__configure(sphinx_index_addjobj)
+        logging.info("QUERY "+sentence)
         rs = self.client_show.Query(sentence, sphinx_index_addjobj)
+        logging.info("OK")
+
+        print json.dumps(rs)
+
+        logging.info("OK")
 
         results = []
         for ma in rs['matches']:

aore/fias/wordentry.py

@@ -1,5 +1,6 @@
 # -*- coding: utf-8 -*-
 import re
+import logging
 
 
 class WordEntry:
@@ -65,7 +66,7 @@ class WordEntry:
         return "({})".format(" | ".join(self.variations))
 
     def __get_ranks(self):
-        word_len = len(self.word)
+        word_len = len(unicode(self.word))
         sql_qry = "SELECT COUNT(*), NULL FROM \"AOTRIG\" WHERE word LIKE '{}%' AND LENGTH(word) > {} " \
                   "UNION ALL SELECT COUNT(*), NULL FROM \"AOTRIG\" WHERE word='{}' " \
                   "UNION ALL SELECT COUNT(*), MAX(scname) FROM \"SOCRBASE\" WHERE socrname ILIKE '{}'" \

aore/templates/postgre/post_create.sql

@@ -4,4 +4,7 @@ CREATE INDEX "sphinx_ind_livestatus" ON "ADDROBJ" USING btree ("actstatus", "liv
 CREATE INDEX "sphinx_ind_aoguid" ON "ADDROBJ" USING btree ("aoguid");
 CREATE INDEX "SOCRBASE_scname_idx" ON "SOCRBASE" USING btree ("scname");
 CREATE INDEX "SOCRBASE_socrname_idx" ON "SOCRBASE" USING btree ("socrname");
+CREATE INDEX "SOCRBASE_scname_gin_idx" ON "SOCRBASE" USING gin(scname gin_trgm_ops);
+CREATE INDEX "SOCRBASE_socrname_gin_idx" ON "SOCRBASE" USING gin(socrname gin_trgm_ops);
 CREATE INDEX "AOTRIG_word_idx" ON "AOTRIG" USING btree ("word");
+CREATE INDEX "AOTRIG_word_gin_idx" ON "AOTRIG" USING gin(word gin_trgm_ops);

aore/updater/updater.py

@@ -72,7 +72,7 @@ class Updater:
         self.__init_update_entries(updates_generator)
         self.db_handler.pre_update()
 
-        for update_entry in self.updates_generator:
+        for update_entry in self.updalist_generator:
             logging.info("Processing update #{}".format(update_entry['intver']))
             for table_entry in self.tablelist_generator(update_entry['delta_url']):
                 self.process_single_entry(table_entry.operation_type, table_entry)

manage.py

@@ -2,7 +2,7 @@
 
 import optparse
 
-from aore.fias.search import SphinxSearch
+from aore.fias.fiasfactory import FiasFactory
 from aore.miscutils.sphinx import SphinxHelper
 from aore.updater.updater import Updater
 from aore.updater.soapreceiver import SoapReceiver
@@ -116,8 +116,8 @@ def main():
 
     # 4 Debug purposes..
     if options.test:
-        sph = SphinxSearch()
+        sph = FiasFactory()
-        sph.find('кедровая пасраул')
+        sph.find('ул кемровая пасраул алтай майминский р-н')
 
 
 if __name__ == '__main__':