Попытка сделать нормальный поиск

README.md

@@ -1,2 +1,11 @@
 # py-fias
-Simple fias app written in Python 2.7
+WSGI application they can serve FIAS (Russian Address Object DB)
+Простое приложение для работы с БД ФИАС, написано для Python 2.7
+
+## Установка
+Протестирована работа на следующих ОС: Windows (8.1) и Debian Jessie
+
+Предполагается, что у Вас уже установлена БД PostgreSql версии 9.5, интерпретатор Python 2.7
+
+1. Windows
+    1. Установим *уйню...

aore/config.py

@@ -37,7 +37,8 @@ SPHINX_VAR_DIRS = dict(
 # config_type = "test"
 
 # Main section
-sphinx_index_name="sph_addrobj"
+sphinx_index_addjobj="idx_fias_addrobj"
+sphinx_index_sugg="idx_fias_sugg"
 sphinx_var_dir=SPHINX_VAR_DIRS[config_type]
 db = DB_INSTANCES[config_type]
 unrar = UNRAR_PATHES[config_type]

aore/fias/search.py

@@ -0,0 +1,20 @@
+# -*- coding: utf-8 -*-
+
+import sphinxapi
+
+import logging
+import json
+
+
+class SphinxSearch:
+    def __init__(self):
+        self.client = sphinxapi.SphinxClient()
+        self.client.SetServer("localhost", 9312)
+        self.client.SetLimits(0, 10)
+
+    def find(self, text):
+        # TODO: ADD index
+        logging.info("12")
+        result = self.client.Query(text)
+        print json.dumps(result)
+        logging.info("12")

aore/miscutils/sphinx.py

@@ -5,18 +5,59 @@ import os
 
 from bottle import template
 
-from aore.config import db as dbconfig, sphinx_index_name, sphinx_var_dir
+from aore.config import db as dbconfig, sphinx_index_addjobj, sphinx_var_dir, trashfolder
 
 
-def produce_sphinx_config(config_name):
-    logging.info("Creating {}".format(config_name))
-    conf_data = template('aore/templates/sphinx/data.conf', db_host=dbconfig['host'], db_user=dbconfig['user'],
+def configure_sphinx(indexer_binary):
+    logging.info("Start configuring Sphinx...")
+
+    # Create ADDROBJ config
+    addrobj_cfg_name = get_addrobj_config()
+
+    # Indexing it...
+    run_index_cmd = "{} -c {} --all".format(indexer_binary, addrobj_cfg_name)
+    logging.info("Run indexer (indexing ADDROBJ)...")
+    os.system(run_index_cmd)
+    logging.info("{} index was created.".format(sphinx_index_addjobj))
+
+    # Produce dict file
+    sugg_dict_name = get_suggestion_dict(indexer_binary, addrobj_cfg_name)
+
+
+def get_suggestion_dict(indexer_binary, addrobj_cfg_name):
+    logging.info("Make suggestion dict...")
+    dict_file_name = os.path.abspath(trashfolder + "suggdict.txt")
+    run_builddict_cmd = "{} {} -c {} --buildstops {} 200000 --buildfreqs".format(indexer_binary, sphinx_index_addjobj,
+                                                                                 addrobj_cfg_name, dict_file_name)
+    os.system(run_builddict_cmd)
+    logging.info("Done.")
+
+    return dict_file_name
+
+
+def get_addrobj_config():
+    config_fname = os.path.abspath(trashfolder + "addrobj.conf")
+    logging.info("Creating config {}".format(config_fname))
+
+    conf_data = template('aore/templates/sphinx/idx_addrobj.conf', db_host=dbconfig['host'], db_user=dbconfig['user'],
                          db_password=dbconfig['password'],
                          db_name=dbconfig['database'], db_port=dbconfig['port'],
-                         sql_query=template('aore/templates/postgre/sphinx_query.sql').replace("\n"," \\\n"), index_name=sphinx_index_name,
+                         sql_query=template('aore/templates/postgre/sphinx_query.sql').replace("\n", " \\\n"),
+                         index_name=sphinx_index_addjobj,
                          sphinx_var_path=sphinx_var_dir)
 
-    conf_data += "\n" + template('aore/templates/sphinx/sphinx.conf', sphinx_var_path=sphinx_var_dir)
+    f = open(config_fname, "w")
+    f.write(conf_data)
+    f.close()
+
+    logging.info("Done.")
+
+    return config_fname
+
+
+# TRASH
+def produce_sphinx_config(config_name):
+    conf_data = template('aore/templates/sphinx/sphinx.conf', sphinx_var_path=sphinx_var_dir)
 
     if os.path.isfile(config_name):
         choice = raw_input(

aore/templates/sphinx/idx_addrobj.conf

@@ -1,4 +1,4 @@
-source src_{{index_name}}
+source {{index_name}}
 {
     type                = pgsql
     sql_host            = {{db_host}}
@@ -15,14 +15,14 @@ source src_{{index_name}}
     sql_attr_uint       = aolevel
 }
 
-index index_{{ index_name }}
+index {{ index_name }}
 {
     docinfo             = extern
     morphology          = stem_ru
-    min_stemming_len    = 2
+    min_stemming_len    = 3
 
     stopwords           =
-    min_word_len        = 2
+    min_word_len        = 1
     charset_type        = utf-8
     min_prefix_len      = 1
     min_infix_len       = 0
@@ -36,6 +36,6 @@ index index_{{ index_name }}
         U+0401->U+0435, U+0451->U+0435, \
         U+410..U+42F->U+430..U+44F, U+430..U+44F
 
-    source              = src_{{index_name}}
-    path                = {{sphinx_var_path}}/data/index_{{index_name}}
+    source              = {{index_name}}
+    path                = {{sphinx_var_path}}/data/{{index_name}}
 }

manage.py

@@ -3,7 +3,8 @@
 import optparse
 
 from aore.aoutils.aoupdater import AoUpdater
-from aore.miscutils.sphinx import produce_sphinx_config
+from aore.miscutils.sphinx import configure_sphinx
+from aore.fias.search import SphinxSearch
 
 
 def update_base(xml_source, updates_count):
@@ -25,8 +26,11 @@ def main():
                  help="Count of updates to process, only for '--database update' option")
     p.add_option('--source', '-s', default="http",
                  help="Create/update DB from source. Value: \"http\" or absolute path to folder")
-    p.add_option('--sphinx-configure', '-c', action="store", type="string",
-                 help="Get Sphinx config. Value: /path/to/sphinx.conf")
+    p.add_option('--sphinx-configure', '-c', action="store_true", dest="sphinx", default="False",
+                 help="Configure sphinx. Creates sphinx.conf in working direcory")
+    p.add_option('--indexer-path', '-i',
+                 help="Path to sphinx indexer binary. Must be specified for '--sphinx-configure'")
+
     options, arguments = p.parse_args()
 
     if options.database:
@@ -37,9 +41,8 @@ def main():
         if options.database == "update":
             update_base(options.source, int(options.update_count))
 
-    if options.sphinx_configure:
-        produce_sphinx_config(options.sphinx_configure)
-
+    if options.sphinx and options.indexer_path:
+        configure_sphinx(options.indexer_path)
 
 if __name__ == '__main__':
     main()