-
Notifications
You must be signed in to change notification settings - Fork 24.9k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
[CI] org.elasticsearch.search.geo.GeoShapeQueryTests. testQueryRandomGeoCollection failure #37356
Comments
Pinging @elastic/es-analytics-geo |
I have opened https://issues.apache.org/jira/browse/LUCENE-8634 |
I muted the test as it failed again today on |
I unmuted the test and added |
I've added some feedback on this issue at https://issues.apache.org/jira/browse/LUCENE-8634 tldr; subcentimeter polygons are not supported w/ the encoding we use in the WGS84 projection. Polygons this small should be reserved for an alternate projection (e.g., mercator, UTM). This will eventually be supported with a new |
related issue popped up today https://elasticsearch-ci.elastic.co/job/elastic+elasticsearch+6.6+release-tests/209/console
|
That looks like a different issue. The original issue was caused by lucene indexing and this issue is caused by parsing. I will take a look. |
Moves the check for tiny polygons earlier in the test. It turned out that polygons can be so tiny that we cannot even figure out their orientation. Relates to elastic#37356
So. It turned out it is a different issue, but with the same cause.
I opened #39275 to ignore the test if such a small polygon is generated for now. However, I am wondering if we should change |
Moves the check for tiny polygons earlier in the test. It turned out that polygons can be so tiny that we cannot even figure out their orientation. Relates to #37356
Moves the check for tiny polygons earlier in the test. It turned out that polygons can be so tiny that we cannot even figure out their orientation. Relates to #37356
Moves the check for tiny polygons earlier in the test. It turned out that polygons can be so tiny that we cannot even figure out their orientation. Relates to #37356
Moves the check for tiny polygons earlier in the test. It turned out that polygons can be so tiny that we cannot even figure out their orientation. Relates to #37356
Moves the check for tiny polygons earlier in the test. It turned out that polygons can be so tiny that we cannot even figure out their orientation. Relates to #37356
I cannot find any failures since the last fix was merged. Closing. |
I noticed some failures in the test above. For example:
https://elasticsearch-ci.elastic.co/job/elastic+elasticsearch+6.6+multijob-unix-compatibility/os=fedora/46/console
It can be reproduced without adding the method name:
./gradlew :server:unitTest
-Dtests.seed=506AA3004455CABB
-Dtests.class=org.elasticsearch.search.geo.GeoShapeQueryTests
-Dtests.method="testQueryRandomGeoCollection"
-Dtests.security.manager=true
-Dtests.locale=it
-Dtests.timezone=Asia/Vladivostok
-Dcompiler.java=11
-Druntime.java=8
I had a look into it and I could reproduce the issue at Lucene level.
The problem is that we are generating degenerated polygons in one dimension, for example:
polygon ((1.401298464324817E-45 22.0, 1.401298464324817E-45 69.0, 4.8202184588118395E-40 69.0, 4.8202184588118395E-40 22.0, 1.401298464324817E-45 22.0))
Then when indexing this polygon and perform an intersect query using the same one, they do not match. This is probably due to the encoding Lucene is using.
The text was updated successfully, but these errors were encountered: