summaryrefslogtreecommitdiffstats
path: root/searchcore/src/tests/proton/matching
Commit message (Collapse)AuthorAgeFilesLines
* Add low-level benchmark program for search iterators.Geir Storli2024-02-081-3/+3
| | | | | Currently, it can benchmark the following query operators over an attribute vector: Single term, In, WeightedSet, DotProduct, Or.
* make default flow stats more explicitHåvard Pettersen2024-02-061-0/+3
| | | | | | for both simple and complex leafs account for number of inner children in complex leafs account for seek nesting for complex leafs with children
* Add feature flag for allow sorting blueprints by cost estimate instead of ↵Henning Baldersheim2023-12-191-232/+119
| | | | est_hits.
* Remove most of the now void clock indirection.Henning Baldersheim2023-12-152-2/+2
|
* Unify on using reference where possible.Henning Baldersheim2023-12-121-3/+3
|
* Wire in thread bundle to execute info and request context.Henning Baldersheim2023-12-122-9/+12
|
* - Control creation of temporary postinglists during fetchPostings for ↵Henning Baldersheim2023-12-041-3/+3
| | | | non-strict iterators.
* Add InTerm to backend.Tor Egge2023-11-242-0/+3
|
* Add flag for marking phrase always expensive.Henning Baldersheim2023-11-191-1/+33
|
* Merge pull request #29369 from vespa-engine/balder/gc-unused-split-parameterHenning Baldersheim2023-11-171-24/+8
|\ | | | | Fully GC unused parameter as we now always split phrases.
| * Fully GC unused parameter as we now always split phrases.Henning Baldersheim2023-11-171-24/+8
| |
* | If hit_rate is below 1% drop match phase limiting. It has too high fixed ↵Henning Baldersheim2023-11-161-0/+1
|/ | | | cost and will liklely make things worse.
* Sameelement behaves like an and with extra constraints.Henning Baldersheim2023-11-071-4/+4
| | | | So it should behave the sameway during fetchPostings too.
* If match-phase limiting has concluded that a post filter is most efficient,Henning Baldersheim2023-11-071-2/+3
| | | | we must only generate posting lists if it is actually benefiscal. If not the fixed cost is too high.
* - Add maxHits param that can be used to compute a realistic(low) hit rate ↵Henning Baldersheim2023-11-032-9/+11
| | | | when fetching summaries (summary features).
* Update copyrightJon Bratseth2023-10-0929-30/+30
|
* Use DfaTable as default fuzzy matching algorithm for maxEditDistance <= 2.Geir Storli2023-10-031-4/+4
|
* Revert "Use DfaTable as default fuzzy matching algorithm for maxEditDistance ↵Henning Baldersheim2023-10-021-4/+4
| | | | …"
* Use DfaTable as default fuzzy matching algorithm for maxEditDistance <= 2.Geir Storli2023-10-021-4/+4
|
* Wire in doom and let hitrate be a float intHenning Baldersheim2023-09-221-3/+3
|
* Add query property to control fuzzy matching algorithm.Geir Storli2023-09-211-6/+14
|
* Control the auto-adjustment of targetHits in ANN using post-filtering.Geir Storli2023-08-151-8/+13
| | | | | | | | | | | | | | 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.
* Avoid createing the FieldSpec on the fly.Henning Baldersheim2023-07-282-23/+19
|
* - Make the MatchContext value object movable.Henning Baldersheim2023-05-261-4/+3
| | | | - Reduce code visibility.
* Use a smallvector to avoid indirection in most common case, and reserve some ↵Henning Baldersheim2023-05-221-2/+2
| | | | space upfront
* As a single field per blueprint is the common use case, make that optimal ↵Henning Baldersheim2023-05-121-1/+1
| | | | more optimal.
* Wire the metastore read guard all the way and use it, if present.Henning Baldersheim2023-05-091-3/+2
|
* Propagate DocumentMetaStore read gaurd to RequestContext.Henning Baldersheim2023-05-092-3/+3
|
* Export moved config definitions.Tor Egge2023-05-091-1/+0
|
* Move ranking config to configdefinitions module.Tor Egge2023-05-093-7/+10
|
* Revert "Move ranking config to configdefinitions module"Harald Musum2023-05-083-10/+7
|
* Move ranking config to configdefinitions module.Tor Egge2023-05-083-7/+10
|
* Move search::FeatureValues to vespalib::FeatureValues in preparation forTor Egge2023-04-251-1/+2
| | | | extending vdslib::SearchResult.
* Drop ranking assets repo from SearchableDocSubDB.Tor Egge2023-02-031-17/+12
|
* separate profiling depth parametersHåvard Pettersen2023-01-191-1/+1
|
* - Remove the explicit close method in SessionManager. Just do it in ↵Henning Baldersheim2023-01-171-10/+10
| | | | | | | destructor instead. - Destruct instead of close. Works even if it has not been constructed. - Minor code cleanup.
* Only attribute fields can represent virtual fields.Geir Storli2023-01-131-7/+13
|
* Imported attributes can also represent virtual fields.Geir Storli2023-01-131-2/+2
|
* Expose SameElement query terms to ranking.Geir Storli2023-01-127-26/+37
| | | | | | A TermFieldMatchData is allocated per SameElement term, and this is used to signal matching docids in doUnpack() on the SameElement search iterator. This allows using the matches() rank feature on a field (virtual) that is searched using a SameElement term.
* Extract virtual fields in the index environment.Geir Storli2023-01-041-0/+35
| | | | | Fields that are represented by a set of attributes in the backend are considered virtual fields. Currently, this is map or array of struct fields (from the SD file) with struct-field attributes.
* Change from typedef to using in searchcore.Geir Storli2022-12-212-5/+5
|
* explicit default filterHåvard Pettersen2022-10-311-0/+3
|
* To prevent against damages caused by excessive hits/offset param being sent ↵Henning Baldersheim2022-10-031-10/+30
| | | | down, cap it against numdocs.
* Update ↵Henning Baldersheim2022-09-131-1/+1
| | | | | searchcore/src/tests/proton/matching/unpacking_iterators_optimizer/unpacking_iterators_optimizer_test.cpp Co-authored-by: Geir Storli <geirst@yahooinc.com>
* Update ↵Henning Baldersheim2022-09-131-1/+1
| | | | | searchcore/src/tests/proton/matching/unpacking_iterators_optimizer/unpacking_iterators_optimizer_test.cpp Co-authored-by: Geir Storli <geirst@yahooinc.com>
* Disable splitting of same-element until we can also handle that with ↵Henning Baldersheim2022-09-131-57/+49
| | | | matched-elements-only.
* thread bundle now available when calculating the global filterHåvard Pettersen2022-09-122-6/+11
|
* Merge pull request #24011 from vespa-engine/havardpe/global-filter-as-interfaceHenning Baldersheim2022-09-121-7/+6
|\ | | | | GlobalFilter is now an interface
| * GlobalFilter is now an interfaceHåvard Pettersen2022-09-121-7/+6
| | | | | | | | instead of a shared optional BitVector
* | Fetch target active from bucketdb.Henning Baldersheim2022-09-111-54/+55
|/