1 # SPDX-License-Identifier: GPL-3.0-or-later
3 # This file is part of Nominatim. (https://nominatim.org)
5 # Copyright (C) 2024 by the Nominatim developer community.
6 # For a full list of authors see the git log.
8 Common data types and protocols for analysers.
10 from typing import Mapping, List, Any
12 from ...typing import Protocol
13 from ...data.place_name import PlaceName
16 class Analyzer(Protocol):
17 """ The `create()` function of an analysis module needs to return an
18 object that implements the following functions.
21 def get_canonical_id(self, name: PlaceName) -> str:
22 """ Return the canonical form of the given name. The canonical ID must
23 be unique (the same ID must always yield the same variants) and
24 must be a form from which the variants can be derived.
27 name: Extended place name description as prepared by
31 ID string with a canonical form of the name. The string may
32 be empty, when the analyzer cannot analyze the name at all,
33 for example because the character set in use does not match.
36 def compute_variants(self, canonical_id: str) -> List[str]:
37 """ Compute the transliterated spelling variants for the given
41 canonical_id: ID string previously computed with
45 A list of possible spelling variants. All strings must have
46 been transformed with the global normalizer and
47 transliterator ICU rules. Otherwise they cannot be matched
48 against the input by the query frontend.
49 The list may be empty, when there are no useful
50 spelling variants. This may happen when an analyzer only
51 usually outputs additional variants to the canonical spelling
52 and there are no such variants.
56 class AnalysisModule(Protocol):
57 """ The setup of the token analysis is split into two parts:
58 configuration and analyser factory. A token analysis module must
59 therefore implement the two functions here described.
62 def configure(self, rules: Mapping[str, Any],
63 normalizer: Any, transliterator: Any) -> Any:
64 """ Prepare the configuration of the analysis module.
65 This function should prepare all data that can be shared
66 between instances of this analyser.
69 rules: A dictionary with the additional configuration options
70 as specified in the tokenizer configuration.
71 normalizer: an ICU Transliterator with the compiled
72 global normalization rules.
73 transliterator: an ICU Transliterator with the compiled
74 global transliteration rules.
77 A data object with configuration data. This will be handed
78 as is into the `create()` function and may be
79 used freely by the analysis module as needed.
82 def create(self, normalizer: Any, transliterator: Any, config: Any) -> Analyzer:
83 """ Create a new instance of the analyser.
84 A separate instance of the analyser is created for each thread
85 when used in multi-threading context.
88 normalizer: an ICU Transliterator with the compiled normalization
90 transliterator: an ICU Transliterator with the compiled
91 transliteration rules.
92 config: The object that was returned by the call to configure().
95 A new analyzer instance. This must be an object that implements
96 the Analyzer protocol.