This copies all necessary data in the project directory to make
sure the tokenizer remains stable even over updates.
"""
+ assert config.project_dir is not None
module_dir = _install_module(config.DATABASE_MODULE_PATH,
config.lib_dir.module,
config.project_dir / 'module')
def init_from_project(self, config: Configuration) -> None:
""" Initialise the tokenizer from the project directory.
"""
+ assert config.project_dir is not None
+
with connect(self.dsn) as conn:
self.normalization = properties.get_property(conn, DBCFG_NORMALIZATION)
def update_sql_functions(self, config: Configuration) -> None:
""" Reimport the SQL functions for this tokenizer.
"""
+ assert config.project_dir is not None
+
with connect(self.dsn) as conn:
max_word_freq = properties.get_property(conn, DBCFG_MAXWORDFREQ)
modulepath = config.DATABASE_MODULE_PATH or \
This is a special migration function for updating existing databases
to new software versions.
"""
+ assert config.project_dir is not None
+
self.normalization = config.TERM_NORMALIZATION
module_dir = _install_module(config.DATABASE_MODULE_PATH,
config.lib_dir.module,
def add_street(self, conn: Connection, street: str) -> None:
""" Add addr:street match terms.
"""
- def _get_street(name: str) -> List[int]:
+ def _get_street(name: str) -> Optional[str]:
with conn.cursor() as cur:
- return cast(List[int],
+ return cast(Optional[str],
cur.scalar("SELECT word_ids_from_name(%s)::text", (name, )))
tokens = self.cache.streets.get(street, _get_street)
- if tokens:
- self.data['street'] = tokens
+ self.data['street'] = tokens or '{}'
def add_place(self, conn: Connection, place: str) -> None: