1 # SPDX-License-Identifier: GPL-2.0-only
3 # This file is part of Nominatim. (https://nominatim.org)
5 # Copyright (C) 2022 by the Nominatim developer community.
6 # For a full list of authors see the git log.
8 Abstract class defintions 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
14 from nominatim.config import Configuration
15 from nominatim.data.place_info import PlaceInfo
17 # pylint: disable=unnecessary-pass
19 class AbstractAnalyzer(ABC):
20 """ The analyzer provides the functions for analysing names and building
23 Analyzers are instantiated on a per-thread base. Access to global data
24 structures must be synchronised accordingly.
27 def __enter__(self) -> 'AbstractAnalyzer':
31 def __exit__(self, exc_type, exc_value, traceback) -> None:
36 def close(self) -> None:
37 """ Free all resources used by the analyzer.
42 def get_word_token_info(self, words: List[str]) -> List[Tuple[str, str, int]]:
43 """ Return token information for the given list of words.
45 The function is used for testing and debugging only
46 and does not need to be particularly efficient.
49 words: A list of words to look up the tokens for.
50 If a word starts with # it is assumed to be a full name
51 otherwise is a partial term.
54 The function returns the list of all tuples that could be
55 found for the given words. Each list entry is a tuple of
56 (original word, word token, word id).
61 def normalize_postcode(self, postcode: str) -> str:
62 """ Convert the postcode to its standardized form.
64 This function must yield exactly the same result as the SQL function
65 `token_normalized_postcode()`.
68 postcode: The postcode to be normalized.
71 The given postcode after normalization.
76 def update_postcodes_from_db(self) -> None:
77 """ Update the tokenizer's postcode tokens from the current content
78 of the `location_postcode` table.
83 def update_special_phrases(self, phrases: List[Tuple[str, str, str, str]],
84 should_replace: bool) -> None:
85 """ Update the tokenizer's special phrase tokens from the given
86 list of special phrases.
89 phrases: The new list of special phrases. Each entry is
90 a tuple of (phrase, class, type, operator).
91 should_replace: If true, replace the current list of phrases.
92 When false, just add the given phrases to the
93 ones that already exist.
98 def add_country_names(self, country_code: str, names: Dict[str, str]):
99 """ Add the given names to the tokenizer's list of country tokens.
102 country_code: two-letter country code for the country the names
104 names: Dictionary of name type to name.
109 def process_place(self, place: PlaceInfo) -> Any:
110 """ Extract tokens for the given place and compute the
111 information to be handed to the PL/pgSQL processor for building
115 place: Place information retrived from the database.
118 A JSON-serialisable structure that will be handed into
119 the database via the `token_info` field.
124 class AbstractTokenizer(ABC):
125 """ The tokenizer instance is the central instance of the tokenizer in
126 the system. There will only be a single instance of the tokenizer
131 def init_new_db(self, config: Configuration, init_db: bool = True) -> None:
132 """ Set up a new tokenizer for the database.
134 The function should copy all necessary data into the project
135 directory or save it in the property table to make sure that
136 the tokenizer remains stable over updates.
139 config: Read-only object with configuration options.
141 init_db: When set to False, then initialisation of database
142 tables should be skipped. This option is only required for
143 migration purposes and can be savely ignored by custom
146 TODO: can we move the init_db parameter somewhere else?
151 def init_from_project(self, config: Configuration) -> None:
152 """ Initialise the tokenizer from an existing database setup.
154 The function should load all previously saved configuration from
155 the project directory and/or the property table.
158 config: Read-only object with configuration options.
163 def finalize_import(self, config: Configuration) -> None:
164 """ This function is called at the very end of an import when all
165 data has been imported and indexed. The tokenizer may create
166 at this point any additional indexes and data structures needed
170 config: Read-only object with configuration options.
175 def update_sql_functions(self, config: Configuration) -> None:
176 """ Update the SQL part of the tokenizer. This function is called
177 automatically on migrations or may be called explicitly by the
178 user through the `nominatim refresh --functions` command.
180 The tokenizer must only update the code of the tokenizer. The
181 data structures or data itself must not be changed by this function.
184 config: Read-only object with configuration options.
189 def check_database(self, config: Configuration) -> str:
190 """ Check that the database is set up correctly and ready for being
194 config: Read-only object with configuration options.
197 If an issue was found, return an error message with the
198 description of the issue as well as hints for the user on
199 how to resolve the issue. If everything is okay, return `None`.
204 def update_statistics(self) -> None:
205 """ Recompute any tokenizer statistics necessary for efficient lookup.
206 This function is meant to be called from time to time by the user
207 to improve performance. However, the tokenizer must not depend on
208 it to be called in order to work.
213 def update_word_tokens(self) -> None:
214 """ Do house-keeping on the tokenizers internal data structures.
215 Remove unused word tokens, resort data etc.
220 def name_analyzer(self) -> AbstractAnalyzer:
221 """ Create a new analyzer for tokenizing names and queries
222 using this tokinzer. Analyzers are context managers and should
226 with tokenizer.name_analyzer() as analyzer:
230 When used outside the with construct, the caller must ensure to
231 call the close() function before destructing the analyzer.