summaryrefslogtreecommitdiffstats
path: root/config-model/src/main/java/com/yahoo/schema/RankProfile.java
Commit message (Collapse)AuthorAgeFilesLines
* Collections.emptyXXX + Collections.singletonXXX => List/Set/Map.ofHenning Baldersheim2024-04-101-6/+5
|
* allow inputs { query(foo) string }Arne Juul2024-02-221-5/+11
|
* fix semantics for empty feature listsArne Juul2024-01-191-21/+9
| | | | | | * allow FeatureList to parse empty input and return empty list * if an empty feature list block is specified in a rank-profile, trigger that we no longer get the implicit inheritance
* explain consequences in the warningArne Juul2023-11-021-1/+1
|
* validate for array/wset attributes, take 2:Arne Juul2023-11-011-2/+33
| | | | | * ensure attribute(arrayfield).count does not trigger validation * log warning via deploylogger instead of throwing exception
* Revert "validate for array/wset attributes"Harald Musum2023-11-011-22/+2
|
* validate for array/wset attributesArne Juul2023-11-011-2/+22
|
* keep the implicit (hidden) match features in a separate set in RankProfileArne Juul2023-10-261-9/+5
|
* avoid losing inherited match-features when adding implicit onesArne Juul2023-10-261-2/+4
|
* allow configuring normalizersArne Juul2023-10-131-10/+102
|
* Update copyrightJon Bratseth2023-10-091-1/+1
|
* if something is already a match-feature, avoid computing it again in ↵Arne Juul2023-09-181-0/+5
| | | | global-phase
* Control the auto-adjustment of targetHits in ANN using post-filtering.Geir Storli2023-08-151-0/+9
| | | | | | | | | | | | | | When searching the HNSW index in a post-filtering case, targetHits is auto-adjusted in an effort to still expose targetHits hits to first-phase ranking after post-filtering. The following formula is now used to ensure an upper bound of adjustedTargetHits, avoiding that the search in the HNSW index takes too long. adjustedTargetHits = min(targetHits / estimatedHitRatio, targetHits * targetHitsMaxAdjustmentFactor). The target-hits-max-adjustment-factor can be set in a rank profile and overriden per query. The value is in the range [1.0,inf], with the default being 20.0. When setting this to 1.0, auto-adjustment of targetHits is effectively disabled.
* add filtering to hide implicitly added match featuresArne Juul2023-05-111-2/+26
|
* - avoid unintentionally duplicated rank-propertiesArne Juul2023-03-201-1/+20
| | | | | - use more LinkedHashMap for deterministic behavior in unit tests - extend some unit tests
* add special handling of TensorFunctionNode containing Generate functionArne Juul2023-03-141-1/+1
|
* use common utility and constantArne Juul2023-03-021-1/+1
|
* add component modelingArne Juul2023-02-241-0/+3
|
* remove debug printArne Juul2023-02-071-1/+0
|
* add inputs as match featuresArne Juul2023-02-071-0/+5
|
* experimental WIPArne Juul2023-02-071-2/+18
|
* wire global-phase through config-model exactly like second-phaseArne Juul2023-02-031-0/+35
|
* Optimize importsHarald Musum2022-10-241-1/+0
|
* More input tests and better error messageJon Bratseth2022-10-191-5/+5
|
* No configuration of phrase splitting needed.Henning Baldersheim2022-09-301-7/+2
|
* Also add option to turn split optimisation off.Henning Baldersheim2022-08-311-0/+2
|
* Remove option to specify delay with feature flag.Henning Baldersheim2022-08-311-3/+0
|
* Control phrase execution with featureflagHenning Baldersheim2022-08-281-23/+22
|
* onnxModel to onnx in summary/matchfeaturesLester Solbakken2022-06-081-2/+2
|
* Rename the 'searchdefinition' package to 'schema'Jon Bratseth2022-05-191-0/+1526