hans/Nominatim
1
0
Fork 1
mirror of https://github.com/osm-search/Nominatim.git synced 2026-02-16 15:47:58 +00:00
Code Issues Packages Projects Releases Wiki Activity

add support for country prefixes in postcodes

Browse Source
This commit is contained in:
Sarah Hoffmann
2025-03-04 15:18:27 +01:00
parent 921db8bb2f
commit 434fbbfd18
1 changed files with 23 additions and 18 deletions
Download Patch File Download Diff File Expand all files Collapse all files

41
src/nominatim_api/search/postcode_parser.py
View File

@@ -7,7 +7,7 @@
"""
Handling of arbitrary postcode tokens in tokenized query string.
"""
from typing import Tuple, Set
from typing import Tuple, Set, Dict, List
import re
from collections import defaultdict
@@ -31,20 +31,21 @@ class PostcodeParser:
cdata = yaml.safe_load(config.find_config_file('country_settings.yaml')
.read_text(encoding='utf-8'))
unique_patterns = defaultdict(set)
unique_patterns: Dict[str, Dict[str, List[str]]] = {}
for cc, data in cdata.items():
if data.get('postcode'):
pat = data['postcode']['pattern']
pat = data['postcode']['pattern'].replace('d', '[0-9]').replace('l', '[a-z]')
out = data['postcode'].get('output')
unique_patterns[pat.replace('d', '[0-9]').replace('l', '[a-z]')].add(out)
if pat not in unique_patterns:
unique_patterns[pat] = defaultdict(list)
unique_patterns[pat][out].append(cc)
self.global_pattern = re.compile(
'(?:' +
'|'.join(f"(?:{k})" for k in unique_patterns)
+ ')[:, >]')
'(?:(?P<cc>[a-z][a-z])(?P<space>[ -]?))?(?P<pc>(?:(?:'
+ ')|(?:'.join(unique_patterns) + '))[:, >].*)')
self.local_patterns = [(re.compile(f"(?:{k})[:, >]"), v)
for k, v in unique_patterns.items()]
self.local_patterns = [(re.compile(f"{pat}[:, >]"), list(info.items()))
for pat, info in unique_patterns.items()]
def parse(self, query: qmod.QueryStruct) -> Set[Tuple[int, int, str]]:
""" Parse postcodes in the given list of query tokens taking into
@@ -64,18 +65,22 @@ class PostcodeParser:
if word[-1] in ' -' and nodes[i + 3].btype != '`':
word += nodes[i + 3].term_normalized + nodes[i + 3].btype
# Use global pattern to check for presence of any postocde.
m = self.global_pattern.match(word)
# Use global pattern to check for presence of any postcode.
m = self.global_pattern.fullmatch(word)
if m:
# If there was a match, check against each pattern separately
# because multiple patterns might be machting at the end.
cc = m.group('cc')
pc_word = m.group('pc')
cc_spaces = len(m.group('space') or '')
for pattern, info in self.local_patterns:
lm = pattern.match(word)
lm = pattern.match(pc_word)
if lm:
trange = (i, i + sum(c in ' ,-:>' for c in lm.group(0)))
for out in info:
if out:
outcodes.add((*trange, lm.expand(out).upper()))
else:
outcodes.add((*trange, lm.group(0)[:-1].upper()))
trange = (i, i + cc_spaces + sum(c in ' ,-:>' for c in lm.group(0)))
for out, out_ccs in info:
if cc is None or cc in out_ccs:
if out:
outcodes.add((*trange, lm.expand(out).upper()))
else:
outcodes.add((*trange, lm.group(0)[:-1].upper()))
return outcodes