src/nominatim_db/tokenizer/icu_token_analysis.py

   1 # SPDX-License-Identifier: GPL-3.0-or-later
   2 #
   3 # This file is part of Nominatim. (https://nominatim.org)
   4 #
   5 # Copyright (C) 2024 by the Nominatim developer community.
   6 # For a full list of authors see the git log.
   7 """
   8 Container class collecting all components required to transform an OSM name
   9 into a Nominatim token.
  10 """
  11 from typing import Mapping, Optional, TYPE_CHECKING
  12 from icu import Transliterator
  13
  14 from .token_analysis.base import Analyzer
  15
  16 if TYPE_CHECKING:
  17     from typing import Any
  18     from .icu_rule_loader import TokenAnalyzerRule # pylint: disable=cyclic-import
  19
  20 class ICUTokenAnalysis:
  21     """ Container class collecting the transliterators and token analysis
  22         modules for a single Analyser instance.
  23     """
  24
  25     def __init__(self, norm_rules: str, trans_rules: str,
  26                  analysis_rules: Mapping[Optional[str], 'TokenAnalyzerRule']):
  27         self.normalizer = Transliterator.createFromRules("icu_normalization",
  28                                                          norm_rules)
  29         trans_rules += ";[:Space:]+ > ' '"
  30         self.to_ascii = Transliterator.createFromRules("icu_to_ascii",
  31                                                        trans_rules)
  32         self.search = Transliterator.createFromRules("icu_search",
  33                                                      norm_rules + trans_rules)
  34
  35         self.analysis = {name: arules.create(self.normalizer, self.to_ascii)
  36                          for name, arules in analysis_rules.items()}
  37
  38
  39     def get_analyzer(self, name: Optional[str]) -> Analyzer:
  40         """ Return the given named analyzer. If no analyzer with that
  41             name exists, return the default analyzer.
  42         """
  43         return self.analysis.get(name) or self.analysis[None]