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

Merge pull request #2736 from lonvia/reverse-interpolation-index-order

Browse Source 
Change indexing order for interpolations and non-addressable objects
This commit is contained in:
Sarah Hoffmann
2022-06-03 10:42:54 +02:00
committed by GitHub
parent 4b0d9f71e8 cbb4749996
commit a024c7665c
2 changed files with 6 additions and 18 deletions
Download Patch File Download Diff File Expand all files Collapse all files

7
nominatim/indexer/indexer.py
View File

@@ -160,15 +160,12 @@ class Indexer:
minrank, maxrank, self.num_threads) minrank, maxrank, self.num_threads)
with self.tokenizer.name_analyzer() as analyzer: with self.tokenizer.name_analyzer() as analyzer:
for rank in range(max(1, minrank), maxrank): for rank in range(max(1, minrank), maxrank + 1):
self._index(runners.RankRunner(rank, analyzer)) self._index(runners.RankRunner(rank, analyzer), 20 if rank == 30 else 1)
if maxrank == 30: if maxrank == 30:
self._index(runners.RankRunner(0, analyzer)) self._index(runners.RankRunner(0, analyzer))
self._index(runners.InterpolationRunner(analyzer), 20) self._index(runners.InterpolationRunner(analyzer), 20)
self._index(runners.RankRunner(30, analyzer), 20)
else:
self._index(runners.RankRunner(maxrank, analyzer))
def index_postcodes(self): def index_postcodes(self):

17
test/python/indexer/test_indexing.py
View File

@@ -177,25 +177,16 @@ def test_index_all_by_rank(test_db, threads, test_tokenizer):
SELECT count(*) FROM placex p WHERE rank_address > 0 SELECT count(*) FROM placex p WHERE rank_address > 0
AND indexed_date >= (SELECT min(indexed_date) FROM placex o AND indexed_date >= (SELECT min(indexed_date) FROM placex o
WHERE p.rank_address < o.rank_address)""") == 0 WHERE p.rank_address < o.rank_address)""") == 0
# placex rank < 30 objects come before interpolations # placex address ranked objects come before interpolations
assert test_db.scalar( assert test_db.scalar(
"""SELECT count(*) FROM placex WHERE rank_address < 30 """SELECT count(*) FROM placex WHERE rank_address > 0
AND indexed_date > AND indexed_date >
(SELECT min(indexed_date) FROM location_property_osmline)""") == 0 (SELECT min(indexed_date) FROM location_property_osmline)""") == 0
# placex rank = 30 objects come after interpolations # rank 0 comes after all other placex objects
assert test_db.scalar( assert test_db.scalar(
"""SELECT count(*) FROM placex WHERE rank_address = 30 """SELECT count(*) FROM placex WHERE rank_address > 0
AND indexed_date <
(SELECT max(indexed_date) FROM location_property_osmline)""") == 0
# rank 0 comes after rank 29 and before rank 30
assert test_db.scalar(
"""SELECT count(*) FROM placex WHERE rank_address < 30
AND indexed_date > AND indexed_date >
(SELECT min(indexed_date) FROM placex WHERE rank_address = 0)""") == 0 (SELECT min(indexed_date) FROM placex WHERE rank_address = 0)""") == 0
assert test_db.scalar(
"""SELECT count(*) FROM placex WHERE rank_address = 30
AND indexed_date <
(SELECT max(indexed_date) FROM placex WHERE rank_address = 0)""") == 0
@pytest.mark.parametrize("threads", [1, 15]) @pytest.mark.parametrize("threads", [1, 15])