move database check for module to tokenizer

nominatim/tokenizer/legacy_tokenizer.py

@@ -132,6 +132,33 @@ class LegacyTokenizer:
                               modulepath=modulepath)
 
 
+    def check_database(self):
+        """ Check that the tokenizer is set up correctly.
+        """
+        hint = """\
+             The Postgresql extension nominatim.so was not correctly loaded.
+
+             Error: {error}
+
+             Hints:
+             * Check the output of the CMmake/make installation step
+             * Does nominatim.so exist?
+             * Does nominatim.so exist on the database server?
+             * Can nominatim.so be accessed by the database user?
+             """
+        with connect(self.dsn) as conn:
+            with conn.cursor() as cur:
+                try:
+                    out = cur.scalar("SELECT make_standard_name('a')")
+                except psycopg2.Error as err:
+                    return hint.format(error=str(err))
+
+        if out != 'a':
+            return hint.format(error='Unexpected result for make_standard_name()')
+
+        return None
+
+
     def migrate_database(self, config):
         """ Initialise the project directory of an existing database for
             use with this tokenizer.

nominatim/tools/check_database.py

@@ -4,10 +4,9 @@ Collection of functions that check if the database is complete and functional.
 from enum import Enum
 from textwrap import dedent
 
-import psycopg2
-
 from nominatim.db.connection import connect
 from nominatim.errors import UsageError
+from nominatim.tokenizer import factory as tokenizer_factory
 
 CHECKLIST = []
 
@@ -78,8 +77,7 @@ def check_database(config):
 
 
 def _get_indexes(conn):
-    indexes = ['idx_word_word_id',
-               'idx_place_addressline_address_place_id',
+    indexes = ['idx_place_addressline_address_place_id',
                'idx_placex_rank_search',
                'idx_placex_rank_address',
                'idx_placex_parent_place_id',
@@ -149,7 +147,7 @@ def check_placex_table(conn, config):
 
 
 @_check(hint="""placex table has no data. Did the import finish sucessfully?""")
-def check_placex_size(conn, config): # pylint: disable=W0613
+def check_placex_size(conn, _):
     """ Checking for placex content
     """
     with conn.cursor() as cur:
@@ -158,38 +156,30 @@ def check_placex_size(conn, config): # pylint: disable=W0613
     return CheckState.OK if cnt > 0 else CheckState.FATAL
 
 
-@_check(hint="""\
-             The Postgresql extension nominatim.so was not correctly loaded.
-
-             Error: {error}
-
-             Hints:
-             * Check the output of the CMmake/make installation step
-             * Does nominatim.so exist?
-             * Does nominatim.so exist on the database server?
-             * Can nominatim.so be accessed by the database user?
-             """)
-def check_module(conn, config): # pylint: disable=W0613
-    """ Checking that nominatim.so module is installed
+@_check(hint="""{msg}""")
+def check_tokenizer(_, config):
+    """ Checking that tokenizer works
     """
-    with conn.cursor() as cur:
-        try:
-            out = cur.scalar("SELECT make_standard_name('a')")
-        except psycopg2.ProgrammingError as err:
-            return CheckState.FAIL, dict(error=str(err))
+    try:
+        tokenizer = tokenizer_factory.get_tokenizer_for_db(config)
+    except UsageError:
+        return CheckState.FAIL, dict(msg="""\
+            Cannot load tokenizer. Did the import finish sucessfully?""")
 
-        if out != 'a':
-            return CheckState.FAIL, dict(error='Unexpected result for make_standard_name()')
+    result = tokenizer.check_database()
 
+    if result is None:
         return CheckState.OK
 
+    return CheckState.FAIL, dict(msg=result)
+
 
 @_check(hint="""\
              The indexing didn't finish. {count} entries are not yet indexed.
 
              To index the remaining entries, run:   {index_cmd}
              """)
-def check_indexing(conn, config): # pylint: disable=W0613
+def check_indexing(conn, _):
     """ Checking indexing status
     """
     with conn.cursor() as cur:
@@ -198,7 +188,7 @@ def check_indexing(conn, config): # pylint: disable=W0613
     if cnt == 0:
         return CheckState.OK
 
-    if conn.index_exists('idx_word_word_id'):
+    if conn.index_exists('idx_placex_rank_search'):
         # Likely just an interrupted update.
         index_cmd = 'nominatim index'
     else:
@@ -214,7 +204,7 @@ def check_indexing(conn, config): # pylint: disable=W0613
 
              Rerun the index creation with:   nominatim import --continue db-postprocess
              """)
-def check_database_indexes(conn, config): # pylint: disable=W0613
+def check_database_indexes(conn, _):
     """ Checking that database indexes are complete
     """
     missing = []
@@ -236,7 +226,7 @@ def check_database_indexes(conn, config): # pylint: disable=W0613
              Invalid indexes:
                {indexes}
              """)
-def check_database_index_valid(conn, config): # pylint: disable=W0613
+def check_database_index_valid(conn, _):
     """ Checking that all database indexes are valid
     """
     with conn.cursor() as cur:

test/python/test_tools_check_database.py

@@ -43,8 +43,22 @@ def test_check_placex_table_size_bad(temp_db_cursor, temp_db_conn, def_config):
     assert chkdb.check_placex_size(temp_db_conn, def_config) == chkdb.CheckState.FATAL
 
 
-def test_check_module_bad(temp_db_conn, def_config):
-    assert chkdb.check_module(temp_db_conn, def_config) == chkdb.CheckState.FAIL
+def test_check_tokenizer_missing(temp_db_conn, def_config, tmp_path):
+    def_config.project_dir = tmp_path
+    assert chkdb.check_tokenizer(temp_db_conn, def_config) == chkdb.CheckState.FAIL
+
+
+@pytest.mark.parametrize("check_result,state", [(None, chkdb.CheckState.OK),
+                                                ("Something wrong", chkdb.CheckState.FAIL)])
+def test_check_tokenizer(tokenizer_mock, temp_db_conn, def_config, monkeypatch,
+                         check_result, state):
+    class _TestTokenizer:
+        def check_database(self):
+            return check_result
+
+    monkeypatch.setattr(chkdb.tokenizer_factory, 'get_tokenizer_for_db',
+                         lambda *a, **k: _TestTokenizer())
+    assert chkdb.check_tokenizer(temp_db_conn, def_config) == state
 
 
 def test_check_indexing_good(temp_db_cursor, temp_db_conn, def_config):