export more data for the tokenizer name preparation

[nominatim.git] / test / python / test_tokenizer_icu.py

diff --git a/test/python/test_tokenizer_icu.py b/test/python/test_tokenizer_icu.py
index 5ec434b6f4b349902ca743106a9199f1382979bc..bbfc0b120d0a98405fb77aabdc1e48266ae541a2 100644 (file)
--- a/test/python/test_tokenizer_icu.py
+++ b/test/python/test_tokenizer_icu.py
@@ -10,6 +10,8 @@ from nominatim.tokenizer import icu_tokenizer
 from nominatim.tokenizer.icu_name_processor import ICUNameProcessorRules
 from nominatim.tokenizer.icu_rule_loader import ICURuleLoader
 from nominatim.db import properties
 from nominatim.tokenizer.icu_name_processor import ICUNameProcessorRules
 from nominatim.tokenizer.icu_rule_loader import ICURuleLoader
 from nominatim.db import properties

+from nominatim.db.sql_preprocessor import SQLPreprocessor
+from nominatim.indexer.place_info import PlaceInfo

 
 from mock_icu_word_table import MockIcuWordTable
 
 
 from mock_icu_word_table import MockIcuWordTable
 


@@ -67,18 +69,24 @@ def analyzer(tokenizer_factory, test_config, monkeypatch,
 
     def _mk_analyser(norm=("[[:Punctuation:][:Space:]]+ > ' '",), trans=(':: upper()',),
                      variants=('~gasse -> gasse', 'street => st', )):
 
     def _mk_analyser(norm=("[[:Punctuation:][:Space:]]+ > ' '",), trans=(':: upper()',),
                      variants=('~gasse -> gasse', 'street => st', )):

-        cfgfile = tmp_path / 'analyser_test_config.yaml'
-        with cfgfile.open('w') as stream:
-            cfgstr = {'normalization' : list(norm),
-                       'transliteration' : list(trans),
-                       'variants' : [ {'words': list(variants)}]}
-            yaml.dump(cfgstr, stream)
-        tok.naming_rules = ICUNameProcessorRules(loader=ICURuleLoader(cfgfile))
+        cfgstr = {'normalization' : list(norm),
+                   'transliteration' : list(trans),
+                   'variants' : [ {'words': list(variants)}]}
+        tok.naming_rules = ICUNameProcessorRules(loader=ICURuleLoader(cfgstr))

 
         return tok.name_analyzer()
 
     return _mk_analyser
 
 
         return tok.name_analyzer()
 
     return _mk_analyser
 

+@pytest.fixture
+def sql_functions(temp_db_conn, def_config, src_dir):
+    orig_sql = def_config.lib_dir.sql
+    def_config.lib_dir.sql = src_dir / 'lib-sql'
+    sqlproc = SQLPreprocessor(temp_db_conn, def_config)
+    sqlproc.run_sql_file(temp_db_conn, 'functions/utils.sql')
+    sqlproc.run_sql_file(temp_db_conn, 'tokenizer/icu_tokenizer.sql')
+    def_config.lib_dir.sql = orig_sql
+

 
 @pytest.fixture
 def getorcreate_full_word(temp_db_cursor):
 
 @pytest.fixture
 def getorcreate_full_word(temp_db_cursor):


@@ -147,7 +155,6 @@ def test_init_new(tokenizer_factory, test_config, monkeypatch, db_prop):
     tok.init_new_db(test_config)
 
     assert db_prop(icu_tokenizer.DBCFG_TERM_NORMALIZATION) == ':: lower();'
     tok.init_new_db(test_config)
 
     assert db_prop(icu_tokenizer.DBCFG_TERM_NORMALIZATION) == ':: lower();'

-    assert db_prop(icu_tokenizer.DBCFG_MAXWORDFREQ) is not None

 
 
 def test_init_word_table(tokenizer_factory, test_config, place_row, word_table):
 
 
 def test_init_word_table(tokenizer_factory, test_config, place_row, word_table):


@@ -166,7 +173,6 @@ def test_init_word_table(tokenizer_factory, test_config, place_row, word_table):
 
 def test_init_from_project(monkeypatch, test_config, tokenizer_factory):
     monkeypatch.setenv('NOMINATIM_TERM_NORMALIZATION', ':: lower();')
 
 def test_init_from_project(monkeypatch, test_config, tokenizer_factory):
     monkeypatch.setenv('NOMINATIM_TERM_NORMALIZATION', ':: lower();')

-    monkeypatch.setenv('NOMINATIM_MAX_WORD_FREQUENCY', '90300')

     tok = tokenizer_factory()
     tok.init_new_db(test_config)
     monkeypatch.undo()
     tok = tokenizer_factory()
     tok.init_new_db(test_config)
     monkeypatch.undo()


@@ -176,23 +182,18 @@ def test_init_from_project(monkeypatch, test_config, tokenizer_factory):
 
     assert tok.naming_rules is not None
     assert tok.term_normalization == ':: lower();'
 
     assert tok.naming_rules is not None
     assert tok.term_normalization == ':: lower();'

-    assert tok.max_word_frequency == '90300'

 
 
 def test_update_sql_functions(db_prop, temp_db_cursor,
                               tokenizer_factory, test_config, table_factory,
                               monkeypatch):
 
 
 def test_update_sql_functions(db_prop, temp_db_cursor,
                               tokenizer_factory, test_config, table_factory,
                               monkeypatch):

-    monkeypatch.setenv('NOMINATIM_MAX_WORD_FREQUENCY', '1133')

     tok = tokenizer_factory()
     tok.init_new_db(test_config)
     tok = tokenizer_factory()
     tok.init_new_db(test_config)

-    monkeypatch.undo()
-
-    assert db_prop(icu_tokenizer.DBCFG_MAXWORDFREQ) == '1133'

 
     table_factory('test', 'txt TEXT')
 
     func_file = test_config.lib_dir.sql / 'tokenizer' / 'icu_tokenizer.sql'
 
     table_factory('test', 'txt TEXT')
 
     func_file = test_config.lib_dir.sql / 'tokenizer' / 'icu_tokenizer.sql'

-    func_file.write_text("""INSERT INTO test VALUES ('{{max_word_freq}}')""")
+    func_file.write_text("""INSERT INTO test VALUES (1133)""")

 
     tok.update_sql_functions(test_config)
 
 
     tok.update_sql_functions(test_config)
 


@@ -307,7 +308,7 @@ def test_add_country_names_extend(analyzer, word_table):
 class TestPlaceNames:
 
     @pytest.fixture(autouse=True)
 class TestPlaceNames:
 
     @pytest.fixture(autouse=True)

-    def setup(self, analyzer, getorcreate_full_word):
+    def setup(self, analyzer, sql_functions):

         with analyzer() as anl:
             self.analyzer = anl
             yield anl
         with analyzer() as anl:
             self.analyzer = anl
             yield anl


@@ -322,30 +323,41 @@ class TestPlaceNames:
         assert eval(info['names']) == set((t[2] for t in tokens))
 
 
         assert eval(info['names']) == set((t[2] for t in tokens))
 
 

+    def process_named_place(self, names):
+        place = {'name': names}
+
+        return self.analyzer.process_place(PlaceInfo(place))
+
+

     def test_simple_names(self):
     def test_simple_names(self):

-        info = self.analyzer.process_place({'name': {'name': 'Soft bAr', 'ref': '34'}})
+        info = self.process_named_place({'name': 'Soft bAr', 'ref': '34'})

 
         self.expect_name_terms(info, '#Soft bAr', '#34', 'Soft', 'bAr', '34')
 
 
     @pytest.mark.parametrize('sep', [',' , ';'])
     def test_names_with_separator(self, sep):
 
         self.expect_name_terms(info, '#Soft bAr', '#34', 'Soft', 'bAr', '34')
 
 
     @pytest.mark.parametrize('sep', [',' , ';'])
     def test_names_with_separator(self, sep):

-        info = self.analyzer.process_place({'name': {'name': sep.join(('New York', 'Big Apple'))}})
+        info = self.process_named_place({'name': sep.join(('New York', 'Big Apple'))})

 
         self.expect_name_terms(info, '#New York', '#Big Apple',
                                      'new', 'york', 'big', 'apple')
 
 
     def test_full_names_with_bracket(self):
 
         self.expect_name_terms(info, '#New York', '#Big Apple',
                                      'new', 'york', 'big', 'apple')
 
 
     def test_full_names_with_bracket(self):

-        info = self.analyzer.process_place({'name': {'name': 'Houseboat (left)'}})
+        info = self.process_named_place({'name': 'Houseboat (left)'})

 
         self.expect_name_terms(info, '#Houseboat (left)', '#Houseboat',
                                      'houseboat', 'left')
 
 
     def test_country_name(self, word_table):
 
         self.expect_name_terms(info, '#Houseboat (left)', '#Houseboat',
                                      'houseboat', 'left')
 
 
     def test_country_name(self, word_table):

-        info = self.analyzer.process_place({'name': {'name': 'Norge'},
-                                           'country_feature': 'no'})
+        place = PlaceInfo({'name' : {'name': 'Norge'},
+                           'country_code': 'no',
+                           'rank_address': 4,
+                           'class': 'boundary',
+                           'type': 'administrative'})
+
+        info = self.analyzer.process_place(place)

 
         self.expect_name_terms(info, '#norge', 'norge')
         assert word_table.get_country() == {('no', 'NORGE')}
 
         self.expect_name_terms(info, '#norge', 'norge')
         assert word_table.get_country() == {('no', 'NORGE')}


@@ -354,14 +366,14 @@ class TestPlaceNames:
 class TestPlaceAddress:
 
     @pytest.fixture(autouse=True)
 class TestPlaceAddress:
 
     @pytest.fixture(autouse=True)

-    def setup(self, analyzer, getorcreate_full_word):
+    def setup(self, analyzer, sql_functions):

         with analyzer(trans=(":: upper()", "'🜵' > ' '")) as anl:
             self.analyzer = anl
             yield anl
 
 
     def process_address(self, **kwargs):
         with analyzer(trans=(":: upper()", "'🜵' > ' '")) as anl:
             self.analyzer = anl
             yield anl
 
 
     def process_address(self, **kwargs):

-        return self.analyzer.process_place({'address': kwargs})
+        return self.analyzer.process_place(PlaceInfo({'address': kwargs}))

 
 
     def name_token_set(self, *expected_terms):
 
 
     def name_token_set(self, *expected_terms):


@@ -427,7 +439,7 @@ class TestPlaceAddress:
     def test_process_place_street(self):
         info = self.process_address(street='Grand Road')
 
     def test_process_place_street(self):
         info = self.process_address(street='Grand Road')
 

-        assert eval(info['street']) == self.name_token_set('#GRAND ROAD')
+        assert eval(info['street']) == self.name_token_set('GRAND', 'ROAD')

 
 
     def test_process_place_street_empty(self):
 
 
     def test_process_place_street_empty(self):


@@ -439,16 +451,13 @@ class TestPlaceAddress:
     def test_process_place_place(self):
         info = self.process_address(place='Honu Lulu')
 
     def test_process_place_place(self):
         info = self.process_address(place='Honu Lulu')
 

-        assert eval(info['place_search']) == self.name_token_set('#HONU LULU',
-                                                                 'HONU', 'LULU')
-        assert eval(info['place_match']) == self.name_token_set('#HONU LULU')
+        assert eval(info['place']) == self.name_token_set('HONU', 'LULU')

 
 
     def test_process_place_place_empty(self):
         info = self.process_address(place='🜵')
 
 
 
     def test_process_place_place_empty(self):
         info = self.process_address(place='🜵')
 

-        assert 'place_search' not in info
-        assert 'place_match' not in info
+        assert 'place' not in info

 
 
     def test_process_place_address_terms(self):
 
 
     def test_process_place_address_terms(self):


@@ -456,16 +465,12 @@ class TestPlaceAddress:
                                     suburb='Zwickau', street='Hauptstr',
                                     full='right behind the church')
 
                                     suburb='Zwickau', street='Hauptstr',
                                     full='right behind the church')
 

-        city_full = self.name_token_set('#ZWICKAU')
-        city_all = self.name_token_set('#ZWICKAU', 'ZWICKAU')
-        state_full = self.name_token_set('#SACHSEN')
-        state_all = self.name_token_set('#SACHSEN', 'SACHSEN')
+        city = self.name_token_set('ZWICKAU')
+        state = self.name_token_set('SACHSEN')

 
 

-        result = {k: [eval(v[0]), eval(v[1])] for k,v in info['addr'].items()}
+        result = {k: eval(v) for k,v in info['addr'].items()}

 
 

-        assert result == {'city': [city_all, city_full],
-                          'suburb': [city_all, city_full],
-                          'state': [state_all, state_full]}
+        assert result == {'city': city, 'suburb': city, 'state': state}

 
 
     def test_process_place_address_terms_empty(self):
 
 
     def test_process_place_address_terms_empty(self):