Commit message (Collapse) | Author | Age | Files | Lines | |
---|---|---|---|---|---|
* | - Show what is mismatch in error message. | Henning Baldersheim | 2023-12-01 | 1 | -0/+5 |
| | | | | - Show extra errors not expected. | ||||
* | Update copyright | Jon Bratseth | 2023-10-09 | 4 | -4/+4 |
| | |||||
* | Add query property to control fuzzy matching algorithm. | Geir Storli | 2023-09-21 | 1 | -0/+2 |
| | |||||
* | Control the auto-adjustment of targetHits in ANN using post-filtering. | Geir Storli | 2023-08-15 | 1 | -1/+6 |
| | | | | | | | | | | | | | | 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. | ||||
* | cosmetic changes | Arne Juul | 2023-02-16 | 1 | -2/+2 |
| | |||||
* | Change from typedef to using in searchlib C++ code. | Geir Storli | 2022-12-21 | 1 | -2/+2 |
| | |||||
* | support mapping names in execution profiling output | Håvard Pettersen | 2022-09-06 | 1 | -31/+31 |
| | | | | | use feature descriptions rather than raw feature names drop single quotes from feature descriptions | ||||
* | Make the message a pair of level and string | Henning Baldersheim | 2022-08-16 | 1 | -36/+53 |
| | |||||
* | Stack usage differs with and without valgrind.., and possibly also OS ↵ | Henning Baldersheim | 2022-08-14 | 1 | -38/+51 |
| | | | | version. Allow for both equality and regex error matching. | ||||
* | - Split rank setup verification code and the binary. | Henning Baldersheim | 2022-08-13 | 1 | -13/+56 |
| | | | | - Accumulate errors during ranksetup. Report them at the end, and also verify them in some tests. | ||||
* | Include searchcommon/attribut/config.h yourself and prepare for hiding large ↵ | Henning Baldersheim | 2022-05-25 | 1 | -0/+1 |
| | | | | | | implementation. Add setters to vespalib::GrowStrategy. | ||||
* | added match features to rank setup | Håvard Pettersen | 2021-11-02 | 1 | -0/+22 |
| | |||||
* | execute -> mutate | Henning Baldersheim | 2021-10-26 | 1 | -12/+16 |
| | |||||
* | Consider the execute instructions comming in the rankprofiles config too. | Henning Baldersheim | 2021-10-12 | 1 | -0/+13 |
| | |||||
* | Update 2017 copyright notices. | gjoranv | 2021-10-07 | 4 | -4/+4 |
| | |||||
* | Explicitly instantiate common attribute update member functions. | Tor Egge | 2020-10-05 | 1 | -1/+1 |
| | |||||
* | better rank feature back-traces | Håvard Pettersen | 2020-03-20 | 1 | -1/+13 |
| | | | | | - log as additional lines on original failure message - skip frames a bit after the middle, not at the very end | ||||
* | increase max dependency depth to 256 | Håvard Pettersen | 2020-03-11 | 1 | -2/+2 |
| | |||||
* | Remove files used by old unit test runner. | Tor Egge | 2019-02-19 | 3 | -3/+0 |
| | |||||
* | enable re-using termwise bitvectors in second phase | Håvard Pettersen | 2017-09-20 | 1 | -8/+15 |
| | | | | | detach match data from rank program and refactor match loop re-use bitvector in termwise search if range is the same | ||||
* | Revert "Havardpe/reuse termwise vector in second phase" | Henning Baldersheim | 2017-09-19 | 1 | -15/+8 |
| | |||||
* | MatchData no longer owned by RankProgram | Håvard Pettersen | 2017-09-19 | 1 | -8/+15 |
| | |||||
* | Update copyright headers | Jon Bratseth | 2017-06-14 | 4 | -4/+4 |
| | |||||
* | Revert "Update copyright headers" | Jon Bratseth | 2017-06-14 | 4 | -4/+4 |
| | |||||
* | Update copyright headers | Jon Bratseth | 2017-06-14 | 4 | -4/+4 |
| | |||||
* | Revert "Copyright header" | Jon Bratseth | 2017-06-13 | 4 | -4/+4 |
| | |||||
* | Copyright header | Jon Bratseth | 2017-06-13 | 4 | -4/+4 |
| | |||||
* | Checkpount for removing fastos from tests | Henning Baldersheim | 2017-06-01 | 1 | -1/+0 |
| | |||||
* | Revert "Revert "Fix warnings hidden earlier due to including application ↵ | Henning Baldersheim | 2017-05-03 | 1 | -5/+2 |
| | | | | headers as s…"" | ||||
* | Revert "Fix warnings hidden earlier due to including application headers as ↵ | Henning Baldersheim | 2017-05-03 | 1 | -2/+5 |
| | | | | s…" | ||||
* | Fix warnings hidden earlier due to including application headers as system ↵ | Henning Baldersheim | 2017-05-03 | 1 | -5/+2 |
| | | | | includes | ||||
* | add override in searchlib module | Arne H Juul | 2017-04-12 | 1 | -2/+2 |
| | |||||
* | Deinline destructorsi and use -Winline gcc option. | Henning Baldersheim | 2017-03-09 | 1 | -12/+20 |
| | |||||
* | lazy evaluate ranking expressions | Haavard | 2017-02-23 | 1 | -22/+15 |
| | |||||
* | Allocate feature executors on a stash to get them close together in memory. | Tor Egge | 2016-12-02 | 1 | -4/+6 |
| | |||||
* | Split out the judymultimap template to separate file and instanciate explicit. | Henning Baldersheim | 2016-11-18 | 1 | -0/+1 |
| | |||||
* | In order to reduce overhead we look at the hit estimate before fanning out ↵ | Henning Baldersheim | 2016-07-06 | 1 | -0/+2 |
| | | | | | | | | to multiple threads. Default is as today, configurable per rank-profile via min-hits-per-thread. Default should be changed to a sane number once verified. The main intentions is to counter the sometimes high cost of per thread rank-setup. | ||||
* | Declared test executables as tests with TEST parameter | Vegard Sjonfjell | 2016-06-16 | 2 | -2/+2 |
| | |||||
* | Publish | Jon Bratseth | 2016-06-15 | 9 | -0/+1005 |