1 # SPDX-License-Identifier: GPL-3.0-or-later
3 # This file is part of Nominatim. (https://nominatim.org)
5 # Copyright (C) 2025 by the Nominatim developer community.
6 # For a full list of authors see the git log.
8 Abstract class definitions for tokenizers. These base classes are here
9 mainly for documentation purposes.
11 from abc import ABC, abstractmethod
12 from typing import List, Tuple, Dict, Any, Optional, Iterable
14 from ..typing import Protocol
15 from ..config import Configuration
16 from ..db.connection import Connection
17 from ..data.place_info import PlaceInfo
20 class AbstractAnalyzer(ABC):
21 """ The analyzer provides the functions for analysing names and building
24 Analyzers are instantiated on a per-thread base. Access to global data
25 structures must be synchronised accordingly.
28 def __enter__(self) -> 'AbstractAnalyzer':
31 def __exit__(self, exc_type: Any, exc_value: Any, traceback: Any) -> None:
35 def close(self) -> None:
36 """ Free all resources used by the analyzer.
40 def get_word_token_info(self, words: List[str]) -> List[Tuple[str, str, int]]:
41 """ Return token information for the given list of words.
43 The function is used for testing and debugging only
44 and does not need to be particularly efficient.
47 words: A list of words to look up the tokens for.
48 If a word starts with # it is assumed to be a full name
49 otherwise is a partial term.
52 The function returns the list of all tuples that could be
53 found for the given words. Each list entry is a tuple of
54 (original word, word token, word id).
58 def normalize_postcode(self, postcode: str) -> str:
59 """ Convert the postcode to its standardized form.
61 This function must yield exactly the same result as the SQL function
62 `token_normalized_postcode()`.
65 postcode: The postcode to be normalized.
68 The given postcode after normalization.
72 def update_postcodes_from_db(self) -> None:
73 """ Update the tokenizer's postcode tokens from the current content
74 of the `location_postcode` table.
78 def update_special_phrases(self,
79 phrases: Iterable[Tuple[str, str, str, str]],
80 should_replace: bool) -> None:
81 """ Update the tokenizer's special phrase tokens from the given
82 list of special phrases.
85 phrases: The new list of special phrases. Each entry is
86 a tuple of (phrase, class, type, operator).
87 should_replace: If true, replace the current list of phrases.
88 When false, just add the given phrases to the
89 ones that already exist.
93 def add_country_names(self, country_code: str, names: Dict[str, str]) -> None:
94 """ Add the given names to the tokenizer's list of country tokens.
97 country_code: two-letter country code for the country the names
99 names: Dictionary of name type to name.
103 def process_place(self, place: PlaceInfo) -> Any:
104 """ Extract tokens for the given place and compute the
105 information to be handed to the PL/pgSQL processor for building
109 place: Place information retrieved from the database.
112 A JSON-serialisable structure that will be handed into
113 the database via the `token_info` field.
117 class AbstractTokenizer(ABC):
118 """ The tokenizer instance is the central instance of the tokenizer in
119 the system. There will only be a single instance of the tokenizer
124 def init_new_db(self, config: Configuration, init_db: bool = True) -> None:
125 """ Set up a new tokenizer for the database.
127 The function should copy all necessary data into the project
128 directory or save it in the property table to make sure that
129 the tokenizer remains stable over updates.
132 config: Read-only object with configuration options.
134 init_db: When set to False, then initialisation of database
135 tables should be skipped. This option is only required for
136 migration purposes and can be safely ignored by custom
141 def init_from_project(self, config: Configuration) -> None:
142 """ Initialise the tokenizer from an existing database setup.
144 The function should load all previously saved configuration from
145 the project directory and/or the property table.
148 config: Read-only object with configuration options.
152 def finalize_import(self, config: Configuration) -> None:
153 """ This function is called at the very end of an import when all
154 data has been imported and indexed. The tokenizer may create
155 at this point any additional indexes and data structures needed
159 config: Read-only object with configuration options.
163 def update_sql_functions(self, config: Configuration) -> None:
164 """ Update the SQL part of the tokenizer. This function is called
165 automatically on migrations or may be called explicitly by the
166 user through the `nominatim refresh --functions` command.
168 The tokenizer must only update the code of the tokenizer. The
169 data structures or data itself must not be changed by this function.
172 config: Read-only object with configuration options.
176 def check_database(self, config: Configuration) -> Optional[str]:
177 """ Check that the database is set up correctly and ready for being
181 config: Read-only object with configuration options.
184 If an issue was found, return an error message with the
185 description of the issue as well as hints for the user on
186 how to resolve the issue. If everything is okay, return `None`.
190 def update_statistics(self, config: Configuration, threads: int = 1) -> None:
191 """ Recompute any tokenizer statistics necessary for efficient lookup.
192 This function is meant to be called from time to time by the user
193 to improve performance. However, the tokenizer must not depend on
194 it to be called in order to work.
198 def update_word_tokens(self) -> None:
199 """ Do house-keeping on the tokenizers internal data structures.
200 Remove unused word tokens, resort data etc.
204 def name_analyzer(self) -> AbstractAnalyzer:
205 """ Create a new analyzer for tokenizing names and queries
206 using this tokinzer. Analyzers are context managers and should
210 with tokenizer.name_analyzer() as analyzer:
214 When used outside the with construct, the caller must ensure to
215 call the close() function before destructing the analyzer.
219 def most_frequent_words(self, conn: Connection, num: int) -> List[str]:
220 """ Return a list of the most frequent full words in the database.
223 conn: Open connection to the database which may be used to
225 num: Maximum number of words to return.
229 class TokenizerModule(Protocol):
230 """ Interface that must be exported by modules that implement their
234 def create(self, dsn: str) -> AbstractTokenizer:
235 """ Factory for new tokenizers.