always state encoding when opening files in text mode

[nominatim.git] / nominatim / tokenizer / legacy_tokenizer.py

diff --git a/nominatim/tokenizer/legacy_tokenizer.py b/nominatim/tokenizer/legacy_tokenizer.py
index 8957426b353efa7ec17f572e754f7fe47f90022c..97ce6d16644cff6a19369c9db5a7a12af8387078 100644 (file)
--- a/nominatim/tokenizer/legacy_tokenizer.py
+++ b/nominatim/tokenizer/legacy_tokenizer.py
@@ -1,3 +1,9 @@
+# SPDX-License-Identifier: GPL-2.0-only
+#
+# This file is part of Nominatim. (https://nominatim.org)
+#
+# Copyright (C) 2022 by the Nominatim developer community.
+# For a full list of authors see the git log.
 """
 Tokenizer implementing normalisation as used before Nominatim 4.
 """
@@ -101,7 +107,7 @@ class LegacyTokenizer(AbstractTokenizer):
 
         self.normalization = config.TERM_NORMALIZATION
 
-        self._install_php(config)
+        self._install_php(config, overwrite=True)
 
         with connect(self.dsn) as conn:
             _check_module(module_dir, conn)
@@ -113,12 +119,18 @@ class LegacyTokenizer(AbstractTokenizer):
             self._init_db_tables(config)
 
 
-    def init_from_project(self):
+    def init_from_project(self, config):
         """ Initialise the tokenizer from the project directory.
         """
         with connect(self.dsn) as conn:
             self.normalization = properties.get_property(conn, DBCFG_NORMALIZATION)
 
+        if not (config.project_dir / 'module' / 'nominatim.so').exists():
+            _install_module(config.DATABASE_MODULE_PATH,
+                            config.lib_dir.module,
+                            config.project_dir / 'module')
+
+        self._install_php(config, overwrite=False)
 
     def finalize_import(self, config):
         """ Do any required postprocessing to make the tokenizer data ready
@@ -142,7 +154,7 @@ class LegacyTokenizer(AbstractTokenizer):
                               modulepath=modulepath)
 
 
-    def check_database(self):
+    def check_database(self, _):
         """ Check that the tokenizer is set up correctly.
         """
         hint = """\
@@ -186,6 +198,32 @@ class LegacyTokenizer(AbstractTokenizer):
             self._save_config(conn, config)
 
 
+    def update_statistics(self):
+        """ Recompute the frequency of full words.
+        """
+        with connect(self.dsn) as conn:
+            if conn.table_exists('search_name'):
+                with conn.cursor() as cur:
+                    cur.drop_table("word_frequencies")
+                    LOG.info("Computing word frequencies")
+                    cur.execute("""CREATE TEMP TABLE word_frequencies AS
+                                     SELECT unnest(name_vector) as id, count(*)
+                                     FROM search_name GROUP BY id""")
+                    cur.execute("CREATE INDEX ON word_frequencies(id)")
+                    LOG.info("Update word table with recomputed frequencies")
+                    cur.execute("""UPDATE word SET search_name_count = count
+                                   FROM word_frequencies
+                                   WHERE word_token like ' %' and word_id = id""")
+                    cur.drop_table("word_frequencies")
+            conn.commit()
+
+
+    def update_word_tokens(self):
+        """ No house-keeping implemented for the legacy tokenizer.
+        """
+        LOG.info("No tokenizer clean-up available.")
+
+
     def name_analyzer(self):
         """ Create a new analyzer for tokenizing names and queries
             using this tokinzer. Analyzers are context managers and should
@@ -206,16 +244,18 @@ class LegacyTokenizer(AbstractTokenizer):
         return LegacyNameAnalyzer(self.dsn, normalizer)
 
 
-    def _install_php(self, config):
+    def _install_php(self, config, overwrite=True):
         """ Install the php script for the tokenizer.
         """
         php_file = self.data_dir / "tokenizer.php"
-        php_file.write_text(dedent("""\
-            <?php
-            @define('CONST_Max_Word_Frequency', {0.MAX_WORD_FREQUENCY});
-            @define('CONST_Term_Normalization_Rules', "{0.TERM_NORMALIZATION}");
-            require_once('{0.lib_dir.php}/tokenizer/legacy_tokenizer.php');
-            """.format(config)))
+
+        if not php_file.exists() or overwrite:
+            php_file.write_text(dedent("""\
+                <?php
+                @define('CONST_Max_Word_Frequency', {0.MAX_WORD_FREQUENCY});
+                @define('CONST_Term_Normalization_Rules', "{0.TERM_NORMALIZATION}");
+                require_once('{0.lib_dir.php}/tokenizer/legacy_tokenizer.php');
+                """.format(config)), encoding='utf-8')
 
 
     def _init_db_tables(self, config):
@@ -405,16 +445,15 @@ class LegacyNameAnalyzer(AbstractAnalyzer):
         """
         token_info = _TokenInfo(self._cache)
 
-        names = place.get('name')
+        names = place.name
 
         if names:
             token_info.add_names(self.conn, names)
 
-            country_feature = place.get('country_feature')
-            if country_feature and re.fullmatch(r'[A-Za-z][A-Za-z]', country_feature):
-                self.add_country_names(country_feature.lower(), names)
+            if place.is_country():
+                self.add_country_names(place.country_code, names)
 
-        address = place.get('address')
+        address = place.address
         if address:
             self._process_place_address(token_info, address)
 
@@ -484,7 +523,7 @@ class _TokenInfo:
             simple_list = list(set(simple_list))
 
         with conn.cursor() as cur:
-            cur.execute("SELECT (create_housenumbers(%s)).* ", (simple_list, ))
+            cur.execute("SELECT * FROM create_housenumbers(%s)", (simple_list, ))
             self.data['hnr_tokens'], self.data['hnr'] = cur.fetchone()
 
 
@@ -495,7 +534,9 @@ class _TokenInfo:
             with conn.cursor() as cur:
                 return cur.scalar("SELECT word_ids_from_name(%s)::text", (name, ))
 
-        self.data['street'] = self.cache.streets.get(street, _get_street)
+        tokens = self.cache.streets.get(street, _get_street)
+        if tokens:
+            self.data['street'] = tokens
 
 
     def add_place(self, conn, place):
@@ -524,9 +565,12 @@ class _TokenInfo:
 
         tokens = {}
         for key, value in terms:
-            tokens[key] = self.cache.address_terms.get(value, _get_address_term)
+            items = self.cache.address_terms.get(value, _get_address_term)
+            if items[0] or items[1]:
+                tokens[key] = items
 
-        self.data['addr'] = tokens
+        if tokens:
+            self.data['addr'] = tokens
 
 
 class _LRU: