[UIMA-6296] SelectFS fsIterator may yield different result set than asList #85
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
…sList - Add test of consitency of forward-iteration through select results with predicates
…ield-different-result-set-than-asList * master: [NO JIRA] Remove Eclipse Maven Repo
…sList - Add test of consitency of back-to-front-iteration through select results with predicates - Add warning when trying to use following/preceding with a negative shift. This worked but it heavily clashes with the ability to make iteration consistent with the asList behavior. Also, a better alternative to following/preceding with negative shift exists, namely using startAt.
…ield-different-result-set-than-asList * master: [UIMA-6254] Move API report post-analysis script into the build resources [UIMA-6254] Move API report post-analysis script into the build resources [UIMA-6254] Move API report post-analysis script into the build resources [UIMA-6254] Move API report post-analysis script into the build resources [UIMA-6276] Potential memory leak in FSClassRegistry
…sList - FsIterator_limited now limits by move operations and not by get operations - Better FsIterator_limited moveTo operation which aims to respect the limit - Fixed another edge case when doing a select in combination with FSIterators - Improved randomized test suite in SelectFsAssert - Move unit tests in SelectFsTest - Switch from assertCountLimit to the more general assertCount in AnnotationIteratorTest
…sList - Fixed various issues encountered when using the moveTo operation in combination with other operations (moveNext/movePrev) on FSIterators obtained via SelectFS - Added various unit tests - Improved randomized test harness - Added debugging code to FsIterator_subtypes_ordered turned off via a static final boolean constant
…sList - Fixed issues involving the limit operation - Refactord tests and added more tests -
…sList - Fixed issues involving non-ambiguous selection
…sList - Adjust the annotation predicate matrix once more to avoid awkward situations in unit tests that before required an explanatory comment - Also, it is now possible to reduce all predicates to coveredBy and there are new tests that include these reductions and ensure that the actual optimized production implementation of the predicates is aligned with the "axiomatic" predicate definition - If we can define all predicates in terms of coveredBy, that means that we can implement basically all of the SelectFS operations by internally mapping them to a "coveredBy" selection. Doing this would likely reduce the complexity of the SelectFS code quite a bit. It will need to be seen if this reduction of code can be done without any significant loss of performance.
…ld-different-result-set-than-asList * main: [UIMA-6295] CAS transportable Java object not serialised or deserialised with compressed binary [UIMA-6295] CAS transportable Java object not serialised or deserialised with compressed binary [UIMA-6295] CAS transportable Java object not serialised or deserialised with compressed binary [UIMA-6295] CAS transportable Java object not serialised or deserialised with compressed binary [UIMA-6295] CAS transportable Java object not serialised or deserialised with compressed binary [UIMA-6295] CAS transportable Java object not serialised or deserialised with compressed binary [UIMA-6301] Rename "master" branches to "main" [UIMA-6295] CAS transportable Java object not serialised or deserialised with compressed binary
…sList - Added test for basic backwards selection - Slightly refactored SelectFsAssert to better organize the sub-asserts within the randomized tests
|
Jenkins, can you test this please? |
…sList - Added additional tests for backwards selection in conjunction with non-overlapping and with limit
…sList - Added test combining non-overlapping, backwards and limit.
…sList - Added test combining non-overlapping and limit. - More refactoring
…sList - Change behavior of shifted by capping it to 0 if it tries moving beyond the selection boundary - Added tests checking the behavior of shifted for default ambigous selects
…sList - Change behavior of shifted by capping it to 0 if it tries moving beyond the selection boundary - Removed unused fields
…sList - Added test for ambigous + backwards + limited + shifted
…sList - Added test for unambigous + backwards + limited + shifted
…ok-to-Asciidoc' into bugfix/UIMA-6296-SelectFS-fsIterator-may-yield-different-result-set-than-asList * feature/UIMA-6305-Convert-UIMAv3-Users-Guide-from-DocBook-to-Asciidoc: [UIMA-6305] Convert UIMAv3 User's Guide from DocBook to Asciidoc [UIMA-6305] Convert UIMAv3 User's Guide from DocBook to Asciidoc [No Jira] Adjust branch names in GitHub metadata % Conflicts: % uima-doc-v3-users-guide/src/docs/asciidoc/images/version_3_users_guide/annotation_predicates/annotation-relations-definition.png % uima-doc-v3-users-guide/src/docs/asciidoc/images/version_3_users_guide/annotation_predicates/annotation-relations-table.png % uima-doc-v3-users-guide/src/docs/asciidoc/images/version_3_users_guide/annotation_predicates/annotation-relations.png % uima-docbook-v3-users-guide/src/docbook/uv3.annotation_predicates.xml % uima-docbook-v3-users-guide/src/docbook/uv3.select.xml
…sList - Describe a few edge cases in the doc
- Build to the proper output directory
…ld-different-result-set-than-asList * main:
reckart
deleted the
bugfix/UIMA-6296-SelectFS-fsIterator-may-yield-different-result-set-than-asList
branch
reckart
added a commit
that referenced
this pull request
Sep 23, 2024
…iption-does-not-discover-type-priorities [UIMA-6204] createReaderDescription does not discover type priorities
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Selection types
[ ] between-> https://issues.apache.org/jira/browse/UIMA-6304FSIterator operations
Selection variations
[ ] use type-priorities-> https://issues.apache.org/jira/browse/UIMA-6304[ ] order not required-> https://issues.apache.org/jira/browse/UIMA-6304[ ] skipWhenSameBeginEndType-> https://issues.apache.org/jira/browse/UIMA-6304[ ] all views-> https://issues.apache.org/jira/browse/UIMA-6304Potential issues:
backwards(limit(shifted(expected)))limit(shifted(backwards(expected)))[10, 20] [20, 30]andselect().preceding([30, 40]), the result would be[10, 20] [20, 30](in this order). So shift normally skips the result list from the beginning, so one might expect the result ofselect().preceding([30, 40]).shifted(1)to be[20, 30]. However, actually shift skips away from the reference point, so the result ofselect().preceding([30, 40]).shifted(1)is actually[10, 20]. I guess that if we take the view thatshiftedis good for selecting relative to the reference annotation instead of relative to the result list, that behavior is ok.startingFsfield which completely ignored ifboundsUse != none. If used with preceding or following, it effectively replaces the reference annotation from which the following/preceding selection is originating. In particular, it does not allow seeking to a particular annotation in the result of the following/preceding selection!