boilerplate for PHP code of tokenizer

This adds an installation step for PHP code for the tokenizer. The
PHP code is split in two parts. The updateable code is found in
lib-php. The tokenizer installs an additional script in the
project directory which then includes the code from lib-php and
defines all settings that are static to the database. The website
code then always includes the PHP from the project directory.

lib-php/Geocode.php

@@ -8,6 +8,7 @@ require_once(CONST_LibDir.'/ReverseGeocode.php');
 require_once(CONST_LibDir.'/SearchDescription.php');
 require_once(CONST_LibDir.'/SearchContext.php');
 require_once(CONST_LibDir.'/TokenList.php');
+require_once(CONST_TokenizerDir.'/tokenizer.php');
 
 class Geocode
 {

lib-php/admin/query.php

@@ -2,7 +2,6 @@
 @define('CONST_LibDir', dirname(dirname(__FILE__)));
 
 require_once(CONST_LibDir.'/init-cmd.php');
-require_once(CONST_LibDir.'/Geocode.php');
 require_once(CONST_LibDir.'/ParameterParser.php');
 ini_set('memory_limit', '800M');
 
@@ -41,16 +40,16 @@ loadSettings($aCMDResult['project-dir'] ?? getcwd());
 @define('CONST_Default_Language', getSetting('DEFAULT_LANGUAGE', false));
 @define('CONST_Log_DB', getSettingBool('LOG_DB'));
 @define('CONST_Log_File', getSetting('LOG_FILE', false));
-@define('CONST_Max_Word_Frequency', getSetting('MAX_WORD_FREQUENCY'));
 @define('CONST_NoAccessControl', getSettingBool('CORS_NOACCESSCONTROL'));
 @define('CONST_Places_Max_ID_count', getSetting('LOOKUP_MAX_COUNT'));
 @define('CONST_PolygonOutput_MaximumTypes', getSetting('POLYGON_OUTPUT_MAX_TYPES'));
 @define('CONST_Search_BatchMode', getSettingBool('SEARCH_BATCH_MODE'));
 @define('CONST_Search_NameOnlySearchFrequencyThreshold', getSetting('SEARCH_NAME_ONLY_THRESHOLD'));
-@define('CONST_Term_Normalization_Rules', getSetting('TERM_NORMALIZATION'));
 @define('CONST_Use_US_Tiger_Data', getSettingBool('USE_US_TIGER_DATA'));
 @define('CONST_MapIcon_URL', getSetting('MAPICON_URL', false));
+@define('CONST_TokenizerDir', CONST_InstallDir.'/tokenizer');
 
+require_once(CONST_LibDir.'/Geocode.php');
 
 $oDB = new Nominatim\DB;
 $oDB->connect();

lib-php/admin/warm.php

@@ -3,7 +3,6 @@
 
 require_once(CONST_LibDir.'/init-cmd.php');
 require_once(CONST_LibDir.'/log.php');
-require_once(CONST_LibDir.'/Geocode.php');
 require_once(CONST_LibDir.'/PlaceLookup.php');
 require_once(CONST_LibDir.'/ReverseGeocode.php');
 
@@ -26,16 +25,16 @@ loadSettings($aCMDResult['project-dir'] ?? getcwd());
 @define('CONST_Default_Language', getSetting('DEFAULT_LANGUAGE', false));
 @define('CONST_Log_DB', getSettingBool('LOG_DB'));
 @define('CONST_Log_File', getSetting('LOG_FILE', false));
-@define('CONST_Max_Word_Frequency', getSetting('MAX_WORD_FREQUENCY'));
 @define('CONST_NoAccessControl', getSettingBool('CORS_NOACCESSCONTROL'));
 @define('CONST_Places_Max_ID_count', getSetting('LOOKUP_MAX_COUNT'));
 @define('CONST_PolygonOutput_MaximumTypes', getSetting('POLYGON_OUTPUT_MAX_TYPES'));
 @define('CONST_Search_BatchMode', getSettingBool('SEARCH_BATCH_MODE'));
 @define('CONST_Search_NameOnlySearchFrequencyThreshold', getSetting('SEARCH_NAME_ONLY_THRESHOLD'));
-@define('CONST_Term_Normalization_Rules', getSetting('TERM_NORMALIZATION'));
 @define('CONST_Use_US_Tiger_Data', getSettingBool('USE_US_TIGER_DATA'));
 @define('CONST_MapIcon_URL', getSetting('MAPICON_URL', false));
+@define('CONST_TokenizerDir', CONST_InstallDir.'/tokenizer');
 
+require_once(CONST_LibDir.'/Geocode.php');
 
 $oDB = new Nominatim\DB();
 $oDB->connect();

lib-php/tokenizer/legacy_tokenizer.php

@@ -0,0 +1 @@
+<?php

nominatim/tokenizer/factory.py

@@ -54,8 +54,7 @@ def create_tokenizer(config, init_db=True, module_name=None):
     tokenizer_module = _import_tokenizer(module_name)
 
     tokenizer = tokenizer_module.create(config.get_libpq_dsn(), basedir)
-    if init_db:
-        tokenizer.init_new_db(config)
+    tokenizer.init_new_db(config, init_db=init_db)
 
     with connect(config.get_libpq_dsn()) as conn:
         properties.set_property(conn, 'tokenizer', module_name)

nominatim/tokenizer/legacy_tokenizer.py

@@ -5,6 +5,7 @@ from collections import OrderedDict
 import logging
 import re
 import shutil
+from textwrap import dedent
 
 from icu import Transliterator
 import psycopg2
@@ -87,7 +88,7 @@ class LegacyTokenizer:
         self.normalization = None
 
 
-    def init_new_db(self, config):
+    def init_new_db(self, config, init_db=True):
         """ Set up a new tokenizer for the database.
 
             This copies all necessary data in the project directory to make
@@ -99,11 +100,14 @@ class LegacyTokenizer:
 
         self.normalization = config.TERM_NORMALIZATION
 
+        self._install_php(config)
+
         with connect(self.dsn) as conn:
             _check_module(module_dir, conn)
             self._save_config(conn, config)
             conn.commit()
 
+        if init_db:
             self.update_sql_functions(config)
             self._init_db_tables(config)
 
@@ -165,6 +169,18 @@ class LegacyTokenizer:
         return LegacyNameAnalyzer(self.dsn, normalizer)
 
 
+    def _install_php(self, config):
+        """ 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)))
+
+
     def _init_db_tables(self, config):
         """ Set up the word table and fill it with pre-computed word
             frequencies.

nominatim/tools/refresh.py

@@ -104,13 +104,11 @@ PHP_CONST_DEFS = (
     ('Default_Language', 'DEFAULT_LANGUAGE', str),
     ('Log_DB', 'LOG_DB', bool),
     ('Log_File', 'LOG_FILE', str),
-    ('Max_Word_Frequency', 'MAX_WORD_FREQUENCY', int),
     ('NoAccessControl', 'CORS_NOACCESSCONTROL', bool),
     ('Places_Max_ID_count', 'LOOKUP_MAX_COUNT', int),
     ('PolygonOutput_MaximumTypes', 'POLYGON_OUTPUT_MAX_TYPES', int),
     ('Search_BatchMode', 'SEARCH_BATCH_MODE', bool),
     ('Search_NameOnlySearchFrequencyThreshold', 'SEARCH_NAME_ONLY_THRESHOLD', str),
-    ('Term_Normalization_Rules', 'TERM_NORMALIZATION', str),
     ('Use_US_Tiger_Data', 'USE_US_TIGER_DATA', bool),
     ('MapIcon_URL', 'MAPICON_URL', str),
 )
@@ -175,9 +173,11 @@ def setup_website(basedir, config):
 
                       @define('CONST_Debug', $_GET['debug'] ?? false);
                       @define('CONST_LibDir', '{0}');
+                      @define('CONST_TokenizerDir', '{2}');
                       @define('CONST_NominatimVersion', '{1[0]}.{1[1]}.{1[2]}-{1[3]}');
 
-                      """.format(config.lib_dir.php, NOMINATIM_VERSION))
+                      """.format(config.lib_dir.php, NOMINATIM_VERSION,
+                                 config.project_dir / 'tokenizer'))
 
     for php_name, conf_name, var_type in PHP_CONST_DEFS:
         if var_type == bool:

test/bdd/steps/nominatim_environment.py

@@ -10,6 +10,7 @@ sys.path.insert(1, str((Path(__file__) / '..' / '..' / '..' / '..').resolve()))
 from nominatim import cli
 from nominatim.config import Configuration
 from nominatim.tools import refresh
+from nominatim.tokenizer import factory as tokenizer_factory
 from steps.utils import run_script
 
 class NominatimEnvironment:
@@ -179,14 +180,10 @@ class NominatimEnvironment:
         """
         self.write_nominatim_config(self.api_test_db)
 
-        if self.api_db_done:
-            return
-
+        if not self.api_db_done:
             self.api_db_done = True
 
-        if self._reuse_or_drop_db(self.api_test_db):
-            return
-
+            if not self._reuse_or_drop_db(self.api_test_db):
                 testdata = Path('__file__') / '..' / '..' / 'testdb'
                 self.test_env['NOMINATIM_WIKIPEDIA_DATA_PATH'] = str(testdata.resolve())
 
@@ -201,6 +198,8 @@ class NominatimEnvironment:
                     self.db_drop_database(self.api_test_db)
                     raise
 
+        tokenizer_factory.create_tokenizer(self.get_test_config(), init_db=False)
+
 
     def setup_unknown_db(self):
         """ Setup a test against a non-existing database.

test/python/dummy_tokenizer.py

@@ -16,7 +16,7 @@ class DummyTokenizer:
         self.analyser_cache = {}
 
 
-    def init_new_db(self, config):
+    def init_new_db(self, *args, **kwargs):
         assert self.init_state == None
         self.init_state = "new"
 

test/python/test_tokenizer_legacy.py

@@ -36,6 +36,7 @@ def test_config(def_config, tmp_path):
 
 @pytest.fixture
 def tokenizer_factory(dsn, tmp_path, monkeypatch, property_table):
+    (tmp_path / 'tokenizer').mkdir()
 
     def _maker():
         return legacy_tokenizer.create(dsn, tmp_path / 'tokenizer')

test/python/test_tools_refresh_setup_website.py

@@ -26,6 +26,7 @@ def test_script(envdir):
 
 def run_website_script(envdir, config):
     config.lib_dir.php = envdir / 'php'
+    config.project_dir = envdir
     refresh.setup_website(envdir, config)
 
     proc = subprocess.run(['/usr/bin/env', 'php', '-Cq',