forked from hans/Nominatim
remove legacy tokenizer and direct tests
This commit is contained in:
Sarah Hoffmann
@@ -1,241 +0,0 @@
|
||||
# SPDX-License-Identifier: GPL-3.0-or-later
|
||||
#
|
||||
# This file is part of Nominatim. (https://nominatim.org)
|
||||
#
|
||||
# Copyright (C) 2024 by the Nominatim developer community.
|
||||
# For a full list of authors see the git log.
|
||||
"""
|
||||
Tests for query analyzer for legacy tokenizer.
|
||||
"""
|
||||
import pytest
|
||||
import pytest_asyncio
|
||||
|
||||
from nominatim_api import NominatimAPIAsync
|
||||
from nominatim_api.search.query import Phrase, PhraseType, TokenType, BreakType
|
||||
import nominatim_api.search.legacy_tokenizer as tok
|
||||
from nominatim_api.logging import set_log_output, get_and_disable
|
||||
|
||||
|
||||
async def add_word(conn, word_id, word_token, word, count):
|
||||
t = conn.t.meta.tables['word']
|
||||
await conn.execute(t.insert(), {'word_id': word_id,
|
||||
'word_token': word_token,
|
||||
'search_name_count': count,
|
||||
'word': word})
|
||||
|
||||
|
||||
async def add_housenumber(conn, word_id, hnr):
|
||||
t = conn.t.meta.tables['word']
|
||||
await conn.execute(t.insert(), {'word_id': word_id,
|
||||
'word_token': ' ' + hnr,
|
||||
'word': hnr,
|
||||
'class': 'place',
|
||||
'type': 'house'})
|
||||
|
||||
|
||||
async def add_postcode(conn, word_id, postcode):
|
||||
t = conn.t.meta.tables['word']
|
||||
await conn.execute(t.insert(), {'word_id': word_id,
|
||||
'word_token': ' ' + postcode,
|
||||
'word': postcode,
|
||||
'class': 'place',
|
||||
'type': 'postcode'})
|
||||
|
||||
|
||||
async def add_special_term(conn, word_id, word_token, cls, typ, op):
|
||||
t = conn.t.meta.tables['word']
|
||||
await conn.execute(t.insert(), {'word_id': word_id,
|
||||
'word_token': word_token,
|
||||
'word': word_token,
|
||||
'class': cls,
|
||||
'type': typ,
|
||||
'operator': op})
|
||||
|
||||
|
||||
def make_phrase(query):
|
||||
return [Phrase(PhraseType.NONE, s) for s in query.split(',')]
|
||||
|
||||
|
||||
@pytest_asyncio.fixture
|
||||
async def conn(table_factory, temp_db_cursor):
|
||||
""" Create an asynchronous SQLAlchemy engine for the test DB.
|
||||
"""
|
||||
table_factory('nominatim_properties',
|
||||
definition='property TEXT, value TEXT',
|
||||
content=(('tokenizer_maxwordfreq', '10000'), ))
|
||||
table_factory('word',
|
||||
definition="""word_id INT, word_token TEXT, word TEXT,
|
||||
class TEXT, type TEXT, country_code TEXT,
|
||||
search_name_count INT, operator TEXT
|
||||
""")
|
||||
|
||||
temp_db_cursor.execute("""CREATE OR REPLACE FUNCTION make_standard_name(name TEXT)
|
||||
RETURNS TEXT AS $$ SELECT lower(name); $$ LANGUAGE SQL;""")
|
||||
|
||||
async with NominatimAPIAsync() as api:
|
||||
async with api.begin() as conn:
|
||||
yield conn
|
||||
|
||||
|
||||
@pytest.mark.asyncio
|
||||
async def test_empty_phrase(conn):
|
||||
ana = await tok.create_query_analyzer(conn)
|
||||
|
||||
query = await ana.analyze_query([])
|
||||
|
||||
assert len(query.source) == 0
|
||||
assert query.num_token_slots() == 0
|
||||
|
||||
|
||||
@pytest.mark.asyncio
|
||||
async def test_single_phrase_with_unknown_terms(conn):
|
||||
ana = await tok.create_query_analyzer(conn)
|
||||
|
||||
await add_word(conn, 1, 'foo', 'FOO', 3)
|
||||
|
||||
query = await ana.analyze_query(make_phrase('foo BAR'))
|
||||
|
||||
assert len(query.source) == 1
|
||||
assert query.source[0].ptype == PhraseType.NONE
|
||||
assert query.source[0].text == 'foo bar'
|
||||
|
||||
assert query.num_token_slots() == 2
|
||||
assert len(query.nodes[0].starting) == 1
|
||||
assert not query.nodes[1].starting
|
||||
|
||||
|
||||
@pytest.mark.asyncio
|
||||
async def test_multiple_phrases(conn):
|
||||
ana = await tok.create_query_analyzer(conn)
|
||||
|
||||
await add_word(conn, 1, 'one', 'one', 13)
|
||||
await add_word(conn, 2, 'two', 'two', 45)
|
||||
await add_word(conn, 100, 'one two', 'one two', 3)
|
||||
await add_word(conn, 3, 'three', 'three', 4584)
|
||||
|
||||
query = await ana.analyze_query(make_phrase('one two,three'))
|
||||
|
||||
assert len(query.source) == 2
|
||||
|
||||
|
||||
@pytest.mark.asyncio
|
||||
async def test_housenumber_token(conn):
|
||||
ana = await tok.create_query_analyzer(conn)
|
||||
|
||||
await add_housenumber(conn, 556, '45 a')
|
||||
|
||||
query = await ana.analyze_query(make_phrase('45 A'))
|
||||
|
||||
assert query.num_token_slots() == 2
|
||||
assert len(query.nodes[0].starting) == 2
|
||||
|
||||
query.nodes[0].starting.sort(key=lambda tl: tl.end)
|
||||
|
||||
hn1 = query.nodes[0].starting[0]
|
||||
assert hn1.ttype == TokenType.HOUSENUMBER
|
||||
assert hn1.end == 1
|
||||
assert hn1.tokens[0].token == 0
|
||||
|
||||
hn2 = query.nodes[0].starting[1]
|
||||
assert hn2.ttype == TokenType.HOUSENUMBER
|
||||
assert hn2.end == 2
|
||||
assert hn2.tokens[0].token == 556
|
||||
|
||||
|
||||
@pytest.mark.asyncio
|
||||
async def test_postcode_token(conn):
|
||||
ana = await tok.create_query_analyzer(conn)
|
||||
|
||||
await add_postcode(conn, 34, '45ax')
|
||||
|
||||
query = await ana.analyze_query(make_phrase('45AX'))
|
||||
|
||||
assert query.num_token_slots() == 1
|
||||
assert [tl.ttype for tl in query.nodes[0].starting] == [TokenType.POSTCODE]
|
||||
|
||||
|
||||
@pytest.mark.asyncio
|
||||
async def test_partial_tokens(conn):
|
||||
ana = await tok.create_query_analyzer(conn)
|
||||
|
||||
await add_word(conn, 1, ' foo', 'foo', 99)
|
||||
await add_word(conn, 1, 'foo', 'FOO', 99)
|
||||
await add_word(conn, 1, 'bar', 'FOO', 990000)
|
||||
|
||||
query = await ana.analyze_query(make_phrase('foo bar'))
|
||||
|
||||
assert query.num_token_slots() == 2
|
||||
|
||||
first = query.nodes[0].starting
|
||||
first.sort(key=lambda tl: tl.tokens[0].penalty)
|
||||
assert [tl.ttype for tl in first] == [TokenType.WORD, TokenType.PARTIAL]
|
||||
assert all(tl.tokens[0].lookup_word == 'foo' for tl in first)
|
||||
|
||||
second = query.nodes[1].starting
|
||||
assert [tl.ttype for tl in second] == [TokenType.PARTIAL]
|
||||
assert not second[0].tokens[0].is_indexed
|
||||
|
||||
|
||||
@pytest.mark.asyncio
|
||||
@pytest.mark.parametrize('term,order', [('23456', ['POSTCODE', 'HOUSENUMBER', 'WORD', 'PARTIAL']),
|
||||
('3', ['HOUSENUMBER', 'POSTCODE', 'WORD', 'PARTIAL'])
|
||||
])
|
||||
async def test_penalty_postcodes_and_housenumbers(conn, term, order):
|
||||
ana = await tok.create_query_analyzer(conn)
|
||||
|
||||
await add_postcode(conn, 1, term)
|
||||
await add_housenumber(conn, 2, term)
|
||||
await add_word(conn, 3, term, term, 5)
|
||||
await add_word(conn, 4, ' ' + term, term, 1)
|
||||
|
||||
query = await ana.analyze_query(make_phrase(term))
|
||||
|
||||
assert query.num_token_slots() == 1
|
||||
|
||||
torder = [(tl.tokens[0].penalty, tl.ttype.name) for tl in query.nodes[0].starting]
|
||||
torder.sort()
|
||||
|
||||
assert [t[1] for t in torder] == order
|
||||
|
||||
|
||||
@pytest.mark.asyncio
|
||||
async def test_category_words_only_at_beginning(conn):
|
||||
ana = await tok.create_query_analyzer(conn)
|
||||
|
||||
await add_special_term(conn, 1, 'foo', 'amenity', 'restaurant', 'in')
|
||||
await add_word(conn, 2, ' bar', 'BAR', 1)
|
||||
|
||||
query = await ana.analyze_query(make_phrase('foo BAR foo'))
|
||||
|
||||
assert query.num_token_slots() == 3
|
||||
assert len(query.nodes[0].starting) == 1
|
||||
assert query.nodes[0].starting[0].ttype == TokenType.NEAR_ITEM
|
||||
assert not query.nodes[2].starting
|
||||
|
||||
|
||||
@pytest.mark.asyncio
|
||||
async def test_qualifier_words(conn):
|
||||
ana = await tok.create_query_analyzer(conn)
|
||||
|
||||
await add_special_term(conn, 1, 'foo', 'amenity', 'restaurant', '-')
|
||||
await add_word(conn, 2, ' bar', 'w', None)
|
||||
|
||||
query = await ana.analyze_query(make_phrase('foo BAR foo BAR foo'))
|
||||
|
||||
assert query.num_token_slots() == 5
|
||||
assert set(t.ttype for t in query.nodes[0].starting) == {TokenType.NEAR_ITEM, TokenType.QUALIFIER}
|
||||
assert set(t.ttype for t in query.nodes[2].starting) == {TokenType.QUALIFIER}
|
||||
assert set(t.ttype for t in query.nodes[4].starting) == {TokenType.NEAR_ITEM, TokenType.QUALIFIER}
|
||||
|
||||
|
||||
@pytest.mark.asyncio
|
||||
@pytest.mark.parametrize('logtype', ['text', 'html'])
|
||||
async def test_log_output(conn, logtype):
|
||||
ana = await tok.create_query_analyzer(conn)
|
||||
|
||||
await add_word(conn, 1, 'foo', 'FOO', 99)
|
||||
|
||||
set_log_output(logtype)
|
||||
await ana.analyze_query(make_phrase('foo'))
|
||||
|
||||
assert get_and_disable()
|
||||
@@ -1,99 +0,0 @@
|
||||
# SPDX-License-Identifier: GPL-3.0-or-later
|
||||
#
|
||||
# This file is part of Nominatim. (https://nominatim.org)
|
||||
#
|
||||
# Copyright (C) 2024 by the Nominatim developer community.
|
||||
# For a full list of authors see the git log.
|
||||
"""
|
||||
Legacy word table for testing with functions to prefil and test contents
|
||||
of the table.
|
||||
"""
|
||||
from nominatim_db.db.connection import execute_scalar
|
||||
|
||||
class MockLegacyWordTable:
|
||||
""" A word table for testing using legacy word table structure.
|
||||
"""
|
||||
def __init__(self, conn):
|
||||
self.conn = conn
|
||||
with conn.cursor() as cur:
|
||||
cur.execute("""CREATE TABLE word (word_id INTEGER,
|
||||
word_token text,
|
||||
word text,
|
||||
class text,
|
||||
type text,
|
||||
country_code varchar(2),
|
||||
search_name_count INTEGER,
|
||||
operator TEXT)""")
|
||||
|
||||
conn.commit()
|
||||
|
||||
def add_full_word(self, word_id, word, word_token=None):
|
||||
with self.conn.cursor() as cur:
|
||||
cur.execute("""INSERT INTO word (word_id, word_token, word)
|
||||
VALUES (%s, %s, %s)
|
||||
""", (word_id, ' ' + (word_token or word), word))
|
||||
self.conn.commit()
|
||||
|
||||
|
||||
def add_special(self, word_token, word, cls, typ, oper):
|
||||
with self.conn.cursor() as cur:
|
||||
cur.execute("""INSERT INTO word (word_token, word, class, type, operator)
|
||||
VALUES (%s, %s, %s, %s, %s)
|
||||
""", (word_token, word, cls, typ, oper))
|
||||
self.conn.commit()
|
||||
|
||||
|
||||
def add_country(self, country_code, word_token):
|
||||
with self.conn.cursor() as cur:
|
||||
cur.execute("INSERT INTO word (word_token, country_code) VALUES(%s, %s)",
|
||||
(word_token, country_code))
|
||||
self.conn.commit()
|
||||
|
||||
|
||||
def add_postcode(self, word_token, postcode):
|
||||
with self.conn.cursor() as cur:
|
||||
cur.execute("""INSERT INTO word (word_token, word, class, type)
|
||||
VALUES (%s, %s, 'place', 'postcode')
|
||||
""", (word_token, postcode))
|
||||
self.conn.commit()
|
||||
|
||||
|
||||
def count(self):
|
||||
return execute_scalar(self.conn, "SELECT count(*) FROM word")
|
||||
|
||||
|
||||
def count_special(self):
|
||||
return execute_scalar(self.conn, "SELECT count(*) FROM word WHERE class != 'place'")
|
||||
|
||||
|
||||
def get_special(self):
|
||||
with self.conn.cursor() as cur:
|
||||
cur.execute("""SELECT word_token, word, class as cls, type, operator
|
||||
FROM word WHERE class != 'place'""")
|
||||
result = set((tuple(row) for row in cur))
|
||||
assert len(result) == cur.rowcount, "Word table has duplicates."
|
||||
return result
|
||||
|
||||
|
||||
def get_country(self):
|
||||
with self.conn.cursor() as cur:
|
||||
cur.execute("""SELECT country_code, word_token
|
||||
FROM word WHERE country_code is not null""")
|
||||
result = set((tuple(row) for row in cur))
|
||||
assert len(result) == cur.rowcount, "Word table has duplicates."
|
||||
return result
|
||||
|
||||
|
||||
def get_postcodes(self):
|
||||
with self.conn.cursor() as cur:
|
||||
cur.execute("""SELECT word FROM word
|
||||
WHERE class = 'place' and type = 'postcode'""")
|
||||
return set((row[0] for row in cur))
|
||||
|
||||
def get_partial_words(self):
|
||||
with self.conn.cursor() as cur:
|
||||
cur.execute("""SELECT word_token, search_name_count FROM word
|
||||
WHERE class is null and country_code is null
|
||||
and not word_token like ' %'""")
|
||||
return set((tuple(row) for row in cur))
|
||||
|
||||
@@ -1,591 +0,0 @@
|
||||
# SPDX-License-Identifier: GPL-3.0-or-later
|
||||
#
|
||||
# This file is part of Nominatim. (https://nominatim.org)
|
||||
#
|
||||
# Copyright (C) 2024 by the Nominatim developer community.
|
||||
# For a full list of authors see the git log.
|
||||
"""
|
||||
Test for legacy tokenizer.
|
||||
"""
|
||||
import shutil
|
||||
import re
|
||||
|
||||
import pytest
|
||||
|
||||
from nominatim_db.data.place_info import PlaceInfo
|
||||
from nominatim_db.tokenizer import legacy_tokenizer
|
||||
from nominatim_db.db import properties
|
||||
from nominatim_db.errors import UsageError
|
||||
|
||||
from mock_legacy_word_table import MockLegacyWordTable
|
||||
|
||||
# Force use of legacy word table
|
||||
@pytest.fixture
|
||||
def word_table(temp_db_conn):
|
||||
return MockLegacyWordTable(temp_db_conn)
|
||||
|
||||
|
||||
@pytest.fixture
|
||||
def test_config(project_env, tmp_path):
|
||||
module_dir = tmp_path / 'module_src'
|
||||
module_dir.mkdir()
|
||||
(module_dir / 'nominatim.so').write_text('TEST nominatim.so')
|
||||
|
||||
project_env.lib_dir.module = module_dir
|
||||
|
||||
sqldir = tmp_path / 'sql'
|
||||
sqldir.mkdir()
|
||||
(sqldir / 'tokenizer').mkdir()
|
||||
|
||||
# Get the original SQL but replace make_standard_name to avoid module use.
|
||||
init_sql = (project_env.lib_dir.sql / 'tokenizer' / 'legacy_tokenizer.sql').read_text()
|
||||
for fn in ('transliteration', 'gettokenstring'):
|
||||
init_sql = re.sub(f'CREATE OR REPLACE FUNCTION {fn}[^;]*;',
|
||||
'', init_sql, re.DOTALL)
|
||||
init_sql += """
|
||||
CREATE OR REPLACE FUNCTION make_standard_name(name TEXT)
|
||||
RETURNS TEXT AS $$ SELECT lower(name); $$ LANGUAGE SQL;
|
||||
|
||||
"""
|
||||
# Also load util functions. Some are needed by the tokenizer.
|
||||
init_sql += (project_env.lib_dir.sql / 'functions' / 'utils.sql').read_text()
|
||||
(sqldir / 'tokenizer' / 'legacy_tokenizer.sql').write_text(init_sql)
|
||||
|
||||
(sqldir / 'words.sql').write_text("SELECT 'a'")
|
||||
|
||||
shutil.copy(str(project_env.lib_dir.sql / 'tokenizer' / 'legacy_tokenizer_tables.sql'),
|
||||
str(sqldir / 'tokenizer' / 'legacy_tokenizer_tables.sql'))
|
||||
|
||||
project_env.lib_dir.sql = sqldir
|
||||
project_env.lib_dir.data = sqldir
|
||||
|
||||
return project_env
|
||||
|
||||
|
||||
@pytest.fixture
|
||||
def tokenizer_factory(dsn, tmp_path, property_table):
|
||||
(tmp_path / 'tokenizer').mkdir()
|
||||
|
||||
def _maker():
|
||||
return legacy_tokenizer.create(dsn, tmp_path / 'tokenizer')
|
||||
|
||||
return _maker
|
||||
|
||||
|
||||
@pytest.fixture
|
||||
def tokenizer_setup(tokenizer_factory, test_config, monkeypatch, sql_preprocessor):
|
||||
monkeypatch.setattr(legacy_tokenizer, '_check_module', lambda m, c: None)
|
||||
tok = tokenizer_factory()
|
||||
tok.init_new_db(test_config)
|
||||
|
||||
|
||||
@pytest.fixture
|
||||
def analyzer(tokenizer_factory, test_config, monkeypatch, sql_preprocessor,
|
||||
word_table, temp_db_with_extensions, tmp_path):
|
||||
monkeypatch.setattr(legacy_tokenizer, '_check_module', lambda m, c: None)
|
||||
monkeypatch.setenv('NOMINATIM_TERM_NORMALIZATION', ':: lower();')
|
||||
tok = tokenizer_factory()
|
||||
tok.init_new_db(test_config)
|
||||
monkeypatch.undo()
|
||||
|
||||
with tok.name_analyzer() as analyzer:
|
||||
yield analyzer
|
||||
|
||||
|
||||
@pytest.fixture
|
||||
def make_standard_name(temp_db_cursor):
|
||||
temp_db_cursor.execute("""CREATE OR REPLACE FUNCTION make_standard_name(name TEXT)
|
||||
RETURNS TEXT AS $$ SELECT '#' || lower(name) || '#'; $$ LANGUAGE SQL""")
|
||||
|
||||
|
||||
@pytest.fixture
|
||||
def create_postcode_id(temp_db_cursor):
|
||||
temp_db_cursor.execute("""CREATE OR REPLACE FUNCTION create_postcode_id(postcode TEXT)
|
||||
RETURNS BOOLEAN AS $$
|
||||
INSERT INTO word (word_token, word, class, type)
|
||||
VALUES (' ' || postcode, postcode, 'place', 'postcode')
|
||||
RETURNING True;
|
||||
$$ LANGUAGE SQL""")
|
||||
|
||||
|
||||
def test_init_new(tokenizer_factory, test_config, monkeypatch,
|
||||
temp_db_conn, sql_preprocessor):
|
||||
monkeypatch.setenv('NOMINATIM_TERM_NORMALIZATION', 'xxvv')
|
||||
monkeypatch.setattr(legacy_tokenizer, '_check_module', lambda m, c: None)
|
||||
|
||||
tok = tokenizer_factory()
|
||||
tok.init_new_db(test_config)
|
||||
|
||||
assert properties.get_property(temp_db_conn, legacy_tokenizer.DBCFG_NORMALIZATION) == 'xxvv'
|
||||
|
||||
outfile = test_config.project_dir / 'module' / 'nominatim.so'
|
||||
|
||||
assert outfile.exists()
|
||||
assert outfile.read_text() == 'TEST nominatim.so'
|
||||
assert outfile.stat().st_mode == 33261
|
||||
|
||||
|
||||
def test_init_module_load_failed(tokenizer_factory, test_config):
|
||||
tok = tokenizer_factory()
|
||||
|
||||
with pytest.raises(UsageError):
|
||||
tok.init_new_db(test_config)
|
||||
|
||||
|
||||
def test_init_module_custom(tokenizer_factory, test_config,
|
||||
monkeypatch, tmp_path, sql_preprocessor):
|
||||
module_dir = (tmp_path / 'custom').resolve()
|
||||
module_dir.mkdir()
|
||||
(module_dir/ 'nominatim.so').write_text('CUSTOM nomiantim.so')
|
||||
|
||||
monkeypatch.setenv('NOMINATIM_DATABASE_MODULE_PATH', str(module_dir))
|
||||
monkeypatch.setattr(legacy_tokenizer, '_check_module', lambda m, c: None)
|
||||
|
||||
tok = tokenizer_factory()
|
||||
tok.init_new_db(test_config)
|
||||
|
||||
assert not (test_config.project_dir / 'module').exists()
|
||||
|
||||
|
||||
def test_init_from_project(tokenizer_setup, tokenizer_factory, test_config):
|
||||
tok = tokenizer_factory()
|
||||
|
||||
tok.init_from_project(test_config)
|
||||
|
||||
assert tok.normalization is not None
|
||||
|
||||
|
||||
def test_update_sql_functions(sql_preprocessor, temp_db_conn,
|
||||
tokenizer_factory, test_config, table_factory,
|
||||
monkeypatch, temp_db_cursor):
|
||||
monkeypatch.setenv('NOMINATIM_MAX_WORD_FREQUENCY', '1133')
|
||||
monkeypatch.setattr(legacy_tokenizer, '_check_module', lambda m, c: None)
|
||||
tok = tokenizer_factory()
|
||||
tok.init_new_db(test_config)
|
||||
monkeypatch.undo()
|
||||
|
||||
assert properties.get_property(temp_db_conn, legacy_tokenizer.DBCFG_MAXWORDFREQ) == '1133'
|
||||
|
||||
table_factory('test', 'txt TEXT')
|
||||
|
||||
func_file = test_config.lib_dir.sql / 'tokenizer' / 'legacy_tokenizer.sql'
|
||||
func_file.write_text("""INSERT INTO test VALUES ('{{max_word_freq}}'),
|
||||
('{{modulepath}}')""")
|
||||
|
||||
tok.update_sql_functions(test_config)
|
||||
|
||||
test_content = temp_db_cursor.row_set('SELECT * FROM test')
|
||||
assert test_content == set((('1133', ), (str(test_config.project_dir / 'module'), )))
|
||||
|
||||
|
||||
def test_finalize_import(tokenizer_factory, temp_db_conn,
|
||||
temp_db_cursor, test_config, monkeypatch,
|
||||
sql_preprocessor_cfg):
|
||||
monkeypatch.setattr(legacy_tokenizer, '_check_module', lambda m, c: None)
|
||||
|
||||
func_file = test_config.lib_dir.sql / 'tokenizer' / 'legacy_tokenizer_indices.sql'
|
||||
func_file.write_text("""CREATE FUNCTION test() RETURNS TEXT
|
||||
AS $$ SELECT 'b'::text $$ LANGUAGE SQL""")
|
||||
|
||||
tok = tokenizer_factory()
|
||||
tok.init_new_db(test_config)
|
||||
|
||||
tok.finalize_import(test_config)
|
||||
|
||||
temp_db_cursor.scalar('SELECT test()') == 'b'
|
||||
|
||||
|
||||
def test_migrate_database(tokenizer_factory, test_config, temp_db_conn, monkeypatch):
|
||||
monkeypatch.setattr(legacy_tokenizer, '_check_module', lambda m, c: None)
|
||||
tok = tokenizer_factory()
|
||||
tok.migrate_database(test_config)
|
||||
|
||||
assert properties.get_property(temp_db_conn, legacy_tokenizer.DBCFG_MAXWORDFREQ) is not None
|
||||
assert properties.get_property(temp_db_conn, legacy_tokenizer.DBCFG_NORMALIZATION) is not None
|
||||
|
||||
outfile = test_config.project_dir / 'module' / 'nominatim.so'
|
||||
|
||||
assert outfile.exists()
|
||||
assert outfile.read_text() == 'TEST nominatim.so'
|
||||
assert outfile.stat().st_mode == 33261
|
||||
|
||||
|
||||
def test_check_database(test_config, tokenizer_factory, monkeypatch,
|
||||
temp_db_cursor, sql_preprocessor_cfg):
|
||||
monkeypatch.setattr(legacy_tokenizer, '_check_module', lambda m, c: None)
|
||||
tok = tokenizer_factory()
|
||||
tok.init_new_db(test_config)
|
||||
|
||||
assert tok.check_database(False) is None
|
||||
|
||||
|
||||
def test_check_database_no_tokenizer(test_config, tokenizer_factory):
|
||||
tok = tokenizer_factory()
|
||||
|
||||
assert tok.check_database(False) is not None
|
||||
|
||||
|
||||
def test_check_database_bad_setup(test_config, tokenizer_factory, monkeypatch,
|
||||
temp_db_cursor, sql_preprocessor_cfg):
|
||||
monkeypatch.setattr(legacy_tokenizer, '_check_module', lambda m, c: None)
|
||||
tok = tokenizer_factory()
|
||||
tok.init_new_db(test_config)
|
||||
|
||||
# Inject a bad transliteration.
|
||||
temp_db_cursor.execute("""CREATE OR REPLACE FUNCTION make_standard_name(name TEXT)
|
||||
RETURNS TEXT AS $$ SELECT 'garbage'::text; $$ LANGUAGE SQL""")
|
||||
|
||||
assert tok.check_database(False) is not None
|
||||
|
||||
|
||||
def test_update_statistics_reverse_only(word_table, tokenizer_factory, test_config):
|
||||
tok = tokenizer_factory()
|
||||
tok.update_statistics(test_config)
|
||||
|
||||
|
||||
def test_update_statistics(word_table, table_factory, temp_db_cursor, tokenizer_factory, test_config):
|
||||
word_table.add_full_word(1000, 'hello')
|
||||
table_factory('search_name',
|
||||
'place_id BIGINT, name_vector INT[]',
|
||||
[(12, [1000])])
|
||||
tok = tokenizer_factory()
|
||||
|
||||
tok.update_statistics(test_config)
|
||||
|
||||
assert temp_db_cursor.scalar("""SELECT count(*) FROM word
|
||||
WHERE word_token like ' %' and
|
||||
search_name_count > 0""") > 0
|
||||
|
||||
|
||||
def test_update_word_tokens(tokenizer_factory):
|
||||
tok = tokenizer_factory()
|
||||
|
||||
# This is a noop and should just pass.
|
||||
tok.update_word_tokens()
|
||||
|
||||
|
||||
def test_normalize(analyzer):
|
||||
assert analyzer.normalize('TEsT') == 'test'
|
||||
|
||||
|
||||
def test_update_postcodes_from_db_empty(analyzer, table_factory, word_table,
|
||||
create_postcode_id):
|
||||
table_factory('location_postcode', 'postcode TEXT',
|
||||
content=(('1234',), ('12 34',), ('AB23',), ('1234',)))
|
||||
|
||||
analyzer.update_postcodes_from_db()
|
||||
|
||||
assert word_table.get_postcodes() == {'1234', '12 34', 'AB23'}
|
||||
|
||||
|
||||
def test_update_postcodes_from_db_add_and_remove(analyzer, table_factory, word_table,
|
||||
create_postcode_id):
|
||||
table_factory('location_postcode', 'postcode TEXT',
|
||||
content=(('1234',), ('45BC', ), ('XX45', )))
|
||||
word_table.add_postcode(' 1234', '1234')
|
||||
word_table.add_postcode(' 5678', '5678')
|
||||
|
||||
analyzer.update_postcodes_from_db()
|
||||
|
||||
assert word_table.get_postcodes() == {'1234', '45BC', 'XX45'}
|
||||
|
||||
|
||||
def test_update_special_phrase_empty_table(analyzer, word_table, make_standard_name):
|
||||
analyzer.update_special_phrases([
|
||||
("König bei", "amenity", "royal", "near"),
|
||||
("Könige", "amenity", "royal", "-"),
|
||||
("könige", "amenity", "royal", "-"),
|
||||
("strasse", "highway", "primary", "in")
|
||||
], True)
|
||||
|
||||
assert word_table.get_special() \
|
||||
== set(((' #könig bei#', 'könig bei', 'amenity', 'royal', 'near'),
|
||||
(' #könige#', 'könige', 'amenity', 'royal', None),
|
||||
(' #strasse#', 'strasse', 'highway', 'primary', 'in')))
|
||||
|
||||
|
||||
def test_update_special_phrase_delete_all(analyzer, word_table, make_standard_name):
|
||||
word_table.add_special(' #foo#', 'foo', 'amenity', 'prison', 'in')
|
||||
word_table.add_special(' #bar#', 'bar', 'highway', 'road', None)
|
||||
|
||||
assert word_table.count_special() == 2
|
||||
|
||||
analyzer.update_special_phrases([], True)
|
||||
|
||||
assert word_table.count_special() == 0
|
||||
|
||||
|
||||
def test_update_special_phrases_no_replace(analyzer, word_table, make_standard_name):
|
||||
word_table.add_special(' #foo#', 'foo', 'amenity', 'prison', 'in')
|
||||
word_table.add_special(' #bar#', 'bar', 'highway', 'road', None)
|
||||
|
||||
assert word_table.count_special() == 2
|
||||
|
||||
analyzer.update_special_phrases([], False)
|
||||
|
||||
assert word_table.count_special() == 2
|
||||
|
||||
|
||||
def test_update_special_phrase_modify(analyzer, word_table, make_standard_name):
|
||||
word_table.add_special(' #foo#', 'foo', 'amenity', 'prison', 'in')
|
||||
word_table.add_special(' #bar#', 'bar', 'highway', 'road', None)
|
||||
|
||||
assert word_table.count_special() == 2
|
||||
|
||||
analyzer.update_special_phrases([
|
||||
('prison', 'amenity', 'prison', 'in'),
|
||||
('bar', 'highway', 'road', '-'),
|
||||
('garden', 'leisure', 'garden', 'near')
|
||||
], True)
|
||||
|
||||
assert word_table.get_special() \
|
||||
== set(((' #prison#', 'prison', 'amenity', 'prison', 'in'),
|
||||
(' #bar#', 'bar', 'highway', 'road', None),
|
||||
(' #garden#', 'garden', 'leisure', 'garden', 'near')))
|
||||
|
||||
|
||||
def test_add_country_names(analyzer, word_table, make_standard_name):
|
||||
analyzer.add_country_names('de', {'name': 'Germany',
|
||||
'name:de': 'Deutschland',
|
||||
'short_name': 'germany'})
|
||||
|
||||
assert word_table.get_country() \
|
||||
== {('de', ' #germany#'),
|
||||
('de', ' #deutschland#')}
|
||||
|
||||
|
||||
def test_add_more_country_names(analyzer, word_table, make_standard_name):
|
||||
word_table.add_country('fr', ' #france#')
|
||||
word_table.add_country('it', ' #italy#')
|
||||
word_table.add_country('it', ' #itala#')
|
||||
|
||||
analyzer.add_country_names('it', {'name': 'Italy', 'ref': 'IT'})
|
||||
|
||||
assert word_table.get_country() \
|
||||
== {('fr', ' #france#'),
|
||||
('it', ' #italy#'),
|
||||
('it', ' #itala#'),
|
||||
('it', ' #it#')}
|
||||
|
||||
|
||||
@pytest.mark.parametrize('pcode', ['12345', 'AB 123', '34-345'])
|
||||
def test_process_place_postcode(analyzer, create_postcode_id, word_table, pcode):
|
||||
analyzer.process_place(PlaceInfo({'address': {'postcode' : pcode}}))
|
||||
|
||||
assert word_table.get_postcodes() == {pcode, }
|
||||
|
||||
|
||||
@pytest.mark.parametrize('pcode', ['12:23', 'ab;cd;f', '123;836'])
|
||||
def test_process_place_bad_postcode(analyzer, create_postcode_id, word_table, pcode):
|
||||
analyzer.process_place(PlaceInfo({'address': {'postcode' : pcode}}))
|
||||
|
||||
assert not word_table.get_postcodes()
|
||||
|
||||
|
||||
class TestHousenumberName:
|
||||
|
||||
@staticmethod
|
||||
@pytest.fixture(autouse=True)
|
||||
def setup_create_housenumbers(temp_db_cursor):
|
||||
temp_db_cursor.execute("""CREATE OR REPLACE FUNCTION create_housenumbers(
|
||||
housenumbers TEXT[],
|
||||
OUT tokens TEXT, OUT normtext TEXT)
|
||||
AS $$
|
||||
SELECT housenumbers::TEXT, array_to_string(housenumbers, ';')
|
||||
$$ LANGUAGE SQL""")
|
||||
|
||||
|
||||
@staticmethod
|
||||
@pytest.mark.parametrize('hnr', ['123a', '1', '101'])
|
||||
def test_process_place_housenumbers_simple(analyzer, hnr):
|
||||
info = analyzer.process_place(PlaceInfo({'address': {'housenumber' : hnr}}))
|
||||
|
||||
assert info['hnr'] == hnr
|
||||
assert info['hnr_tokens'].startswith("{")
|
||||
|
||||
|
||||
@staticmethod
|
||||
def test_process_place_housenumbers_lists(analyzer):
|
||||
info = analyzer.process_place(PlaceInfo({'address': {'conscriptionnumber' : '1; 2;3'}}))
|
||||
|
||||
assert set(info['hnr'].split(';')) == set(('1', '2', '3'))
|
||||
|
||||
|
||||
@staticmethod
|
||||
def test_process_place_housenumbers_duplicates(analyzer):
|
||||
info = analyzer.process_place(PlaceInfo({'address': {'housenumber' : '134',
|
||||
'conscriptionnumber' : '134',
|
||||
'streetnumber' : '99a'}}))
|
||||
|
||||
assert set(info['hnr'].split(';')) == set(('134', '99a'))
|
||||
|
||||
|
||||
class TestPlaceNames:
|
||||
|
||||
@pytest.fixture(autouse=True)
|
||||
def setup(self, analyzer):
|
||||
self.analyzer = analyzer
|
||||
|
||||
|
||||
def expect_name_terms(self, info, *expected_terms):
|
||||
tokens = self.analyzer.get_word_token_info(list(expected_terms))
|
||||
for token in tokens:
|
||||
assert token[2] is not None, "No token for {0}".format(token)
|
||||
|
||||
assert eval(info['names']) == set((t[2] for t in tokens)),\
|
||||
f"Expected: {tokens}\nGot: {info['names']}"
|
||||
|
||||
|
||||
def process_named_place(self, names):
|
||||
return self.analyzer.process_place(PlaceInfo({'name': names}))
|
||||
|
||||
|
||||
def test_simple_names(self):
|
||||
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):
|
||||
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):
|
||||
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):
|
||||
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')}
|
||||
|
||||
|
||||
class TestPlaceAddress:
|
||||
|
||||
@pytest.fixture(autouse=True)
|
||||
def setup(self, analyzer):
|
||||
self.analyzer = analyzer
|
||||
|
||||
|
||||
@pytest.fixture
|
||||
def getorcreate_hnr_id(self, temp_db_cursor):
|
||||
temp_db_cursor.execute("""CREATE SEQUENCE seq_hnr start 1;
|
||||
CREATE OR REPLACE FUNCTION getorcreate_housenumber_id(lookup_word TEXT)
|
||||
RETURNS INTEGER AS $$
|
||||
SELECT -nextval('seq_hnr')::INTEGER; $$ LANGUAGE SQL""")
|
||||
|
||||
def process_address(self, **kwargs):
|
||||
return self.analyzer.process_place(PlaceInfo({'address': kwargs}))
|
||||
|
||||
|
||||
def name_token_set(self, *expected_terms):
|
||||
tokens = self.analyzer.get_word_token_info(list(expected_terms))
|
||||
for token in tokens:
|
||||
assert token[2] is not None, "No token for {0}".format(token)
|
||||
|
||||
return set((t[2] for t in tokens))
|
||||
|
||||
|
||||
@pytest.mark.parametrize('pcode', ['12345', 'AB 123', '34-345'])
|
||||
def test_process_place_postcode(self, word_table, pcode):
|
||||
self.process_address(postcode=pcode)
|
||||
|
||||
assert word_table.get_postcodes() == {pcode, }
|
||||
|
||||
|
||||
@pytest.mark.parametrize('pcode', ['12:23', 'ab;cd;f', '123;836'])
|
||||
def test_process_place_bad_postcode(self, word_table, pcode):
|
||||
self.process_address(postcode=pcode)
|
||||
|
||||
assert not word_table.get_postcodes()
|
||||
|
||||
|
||||
@pytest.mark.parametrize('hnr', ['123a', '0', '101'])
|
||||
def test_process_place_housenumbers_simple(self, hnr, getorcreate_hnr_id):
|
||||
info = self.process_address(housenumber=hnr)
|
||||
|
||||
assert info['hnr'] == hnr.lower()
|
||||
assert info['hnr_tokens'] == "{-1}"
|
||||
|
||||
|
||||
def test_process_place_housenumbers_lists(self, getorcreate_hnr_id):
|
||||
info = self.process_address(conscriptionnumber='1; 2;3')
|
||||
|
||||
assert set(info['hnr'].split(';')) == set(('1', '2', '3'))
|
||||
assert info['hnr_tokens'] == "{-1,-2,-3}"
|
||||
|
||||
|
||||
def test_process_place_housenumbers_duplicates(self, getorcreate_hnr_id):
|
||||
info = self.process_address(housenumber='134',
|
||||
conscriptionnumber='134',
|
||||
streetnumber='99A')
|
||||
|
||||
assert set(info['hnr'].split(';')) == set(('134', '99a'))
|
||||
assert info['hnr_tokens'] == "{-1,-2}"
|
||||
|
||||
|
||||
def test_process_place_street(self):
|
||||
# legacy tokenizer only indexes known names
|
||||
self.analyzer.process_place(PlaceInfo({'name': {'name' : 'Grand Road'}}))
|
||||
info = self.process_address(street='Grand Road')
|
||||
|
||||
assert eval(info['street']) == self.name_token_set('#Grand Road')
|
||||
|
||||
|
||||
def test_process_place_street_empty(self):
|
||||
info = self.process_address(street='🜵')
|
||||
|
||||
assert info['street'] == '{}'
|
||||
|
||||
|
||||
def test_process_place_place(self):
|
||||
self.analyzer.process_place(PlaceInfo({'name': {'name' : 'Honu Lulu'}}))
|
||||
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')
|
||||
|
||||
|
||||
def test_process_place_place_empty(self):
|
||||
info = self.process_address(place='🜵')
|
||||
|
||||
assert 'place' not in info
|
||||
|
||||
|
||||
def test_process_place_address_terms(self):
|
||||
for name in ('Zwickau', 'Haupstraße', 'Sachsen'):
|
||||
self.analyzer.process_place(PlaceInfo({'name': {'name' : name}}))
|
||||
info = self.process_address(country='de', city='Zwickau', state='Sachsen',
|
||||
suburb='Zwickau', street='Hauptstr',
|
||||
full='right behind the church')
|
||||
|
||||
city = self.name_token_set('ZWICKAU')
|
||||
state = self.name_token_set('SACHSEN')
|
||||
|
||||
print(info)
|
||||
result = {k: eval(v[0]) for k,v in info['addr'].items()}
|
||||
|
||||
assert result == {'city': city, 'suburb': city, 'state': state}
|
||||
|
||||
|
||||
def test_process_place_address_terms_empty(self):
|
||||
info = self.process_address(country='de', city=' ', street='Hauptstr',
|
||||
full='right behind the church')
|
||||
|
||||
assert 'addr' not in info
|
||||
|
||||
@@ -14,8 +14,6 @@ from nominatim_db.errors import UsageError
|
||||
from nominatim_db.db.connection import server_version_tuple
|
||||
import nominatim_db.version
|
||||
|
||||
from mock_legacy_word_table import MockLegacyWordTable
|
||||
|
||||
class DummyTokenizer:
|
||||
|
||||
def update_sql_functions(self, config):
|
||||
@@ -28,10 +26,6 @@ def postprocess_mock(monkeypatch):
|
||||
monkeypatch.setattr(migration.tokenizer_factory, 'get_tokenizer_for_db',
|
||||
lambda *args: DummyTokenizer())
|
||||
|
||||
@pytest.fixture
|
||||
def legacy_word_table(temp_db_conn):
|
||||
return MockLegacyWordTable(temp_db_conn)
|
||||
|
||||
|
||||
def test_no_migration_old_versions(temp_db_with_extensions, table_factory, def_config):
|
||||
table_factory('country_name', 'name HSTORE, country_code TEXT')
|
||||
@@ -145,26 +139,6 @@ def test_add_nominatim_property_table_repeat(temp_db_conn, temp_db_cursor,
|
||||
assert temp_db_cursor.table_exists('nominatim_properties')
|
||||
|
||||
|
||||
def test_change_housenumber_transliteration(temp_db_conn, temp_db_cursor,
|
||||
legacy_word_table, placex_table):
|
||||
placex_table.add(housenumber='3A')
|
||||
|
||||
temp_db_cursor.execute("""CREATE OR REPLACE FUNCTION make_standard_name(name TEXT)
|
||||
RETURNS TEXT AS $$ SELECT lower(name) $$ LANGUAGE SQL """)
|
||||
temp_db_cursor.execute("""CREATE OR REPLACE FUNCTION getorcreate_housenumber_id(lookup_word TEXT)
|
||||
RETURNS INTEGER AS $$ SELECT 4325 $$ LANGUAGE SQL """)
|
||||
|
||||
migration.change_housenumber_transliteration(temp_db_conn)
|
||||
temp_db_conn.commit()
|
||||
|
||||
assert temp_db_cursor.scalar('SELECT housenumber from placex') == '3a'
|
||||
|
||||
migration.change_housenumber_transliteration(temp_db_conn)
|
||||
temp_db_conn.commit()
|
||||
|
||||
assert temp_db_cursor.scalar('SELECT housenumber from placex') == '3a'
|
||||
|
||||
|
||||
def test_switch_placenode_geometry_index(temp_db_conn, temp_db_cursor, placex_table):
|
||||
temp_db_cursor.execute("""CREATE INDEX idx_placex_adminname
|
||||
ON placex (place_id)""")
|
||||
|
||||
Reference in New Issue
Block a user