hans/Nominatim
1
0
Fork 1
mirror of https://github.com/osm-search/Nominatim.git synced 2026-02-26 11:08:13 +00:00
Code Issues Packages Projects Releases Wiki Activity

avoid splitting of first token when a housenumber is present

Browse Source 
This only covers the case of <poi name> <street name> <housenumber>
which is exceedingly rare.
This commit is contained in:
Sarah Hoffmann
2023-07-14 21:05:13 +02:00
parent 8a36ed4f6f
commit 283db76e45
2 changed files with 8 additions and 4 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
nominatim/api/search/db_search_builder.py
View File

@@ -225,9 +225,7 @@ class SearchBuilder:
# This might yield wrong results, nothing we can do about that. # This might yield wrong results, nothing we can do about that.
if not partials_indexed: if not partials_indexed:
addr_tokens = [t.token for t in addr_partials if t.is_indexed] addr_tokens = [t.token for t in addr_partials if t.is_indexed]
log().var_dump('before', penalty)
penalty += 1.2 * sum(t.penalty for t in addr_partials if not t.is_indexed) penalty += 1.2 * sum(t.penalty for t in addr_partials if not t.is_indexed)
log().var_dump('after', penalty)
if rare_names: if rare_names:
# Any of the full names applies with all of the partials from the address # Any of the full names applies with all of the partials from the address
lookup = [dbf.FieldLookup('name_vector', [t.token for t in rare_names], 'lookup_any')] lookup = [dbf.FieldLookup('name_vector', [t.token for t in rare_names], 'lookup_any')]

10
nominatim/api/search/token_assignment.py
View File

@@ -309,9 +309,12 @@ class _TokenSequence:
first = base.address[0] first = base.address[0]
if (not base.housenumber or first.end >= base.housenumber.start)\ if (not base.housenumber or first.end >= base.housenumber.start)\
and (not base.qualifier or first.start >= base.qualifier.end): and (not base.qualifier or first.start >= base.qualifier.end):
base_penalty = self.penalty
if base.housenumber and base.housenumber.start > first.start:
base_penalty += 0.25
for i in range(first.start + 1, first.end): for i in range(first.start + 1, first.end):
name, addr = first.split(i) name, addr = first.split(i)
penalty = self.penalty + PENALTY_TOKENCHANGE[query.nodes[i].btype] penalty = base_penalty + PENALTY_TOKENCHANGE[query.nodes[i].btype]
log().comment(f'split first word = name ({i - first.start})') log().comment(f'split first word = name ({i - first.start})')
yield dataclasses.replace(base, name=name, penalty=penalty, yield dataclasses.replace(base, name=name, penalty=penalty,
address=[addr] + base.address[1:]) address=[addr] + base.address[1:])
@@ -321,9 +324,12 @@ class _TokenSequence:
last = base.address[-1] last = base.address[-1]
if (not base.housenumber or last.start <= base.housenumber.end)\ if (not base.housenumber or last.start <= base.housenumber.end)\
and (not base.qualifier or last.end <= base.qualifier.start): and (not base.qualifier or last.end <= base.qualifier.start):
base_penalty = self.penalty
if base.housenumber and base.housenumber.start < last.start:
base_penalty += 0.4
for i in range(last.start + 1, last.end): for i in range(last.start + 1, last.end):
addr, name = last.split(i) addr, name = last.split(i)
penalty = self.penalty + PENALTY_TOKENCHANGE[query.nodes[i].btype] penalty = base_penalty + PENALTY_TOKENCHANGE[query.nodes[i].btype]
log().comment(f'split last word = name ({i - last.start})') log().comment(f'split last word = name ({i - last.start})')
yield dataclasses.replace(base, name=name, penalty=penalty, yield dataclasses.replace(base, name=name, penalty=penalty,
address=base.address[:-1] + [addr]) address=base.address[:-1] + [addr])