Commit message (Collapse) | Author | Age | Files | Lines | |
---|---|---|---|---|---|
* | encode -> embed | Jon Bratseth | 2021-09-28 | 2 | -15/+14 |
| | |||||
* | Encode tensors passed as encode(text) | Jon Bratseth | 2021-09-20 | 1 | -8/+77 |
| | |||||
* | Merge pull request #19003 from takamabe/correct_heuristic_for_negativeItem | Jon Bratseth | 2021-09-10 | 1 | -0/+7 |
|\ | | | | | correct heuristic for -(N) | ||||
| * | correct heuristic for -(N) | takamabe | 2021-09-08 | 1 | -0/+7 |
| | | |||||
* | | set timeout for clone query | takamabe | 2021-08-27 | 1 | -0/+8 |
|/ | |||||
* | we want to compare Linguistics objects for equivalence | Arne Juul | 2021-08-04 | 2 | -0/+6 |
| | |||||
* | Merge pull request #18511 from vespa-engine/bratseth/unbalanced-with-few-docs | Henning Baldersheim | 2021-07-02 | 5 | -60/+88 |
|\ | | | | | Bratseth/unbalanced with few docs | ||||
| * | Enable top-k optimization only if balanced and non-sparse | Jon Bratseth | 2021-07-02 | 1 | -24/+46 |
| | | |||||
| * | Separate balanced and sparse | Jon Bratseth | 2021-07-02 | 1 | -2/+22 |
| | | |||||
| * | Infer group | Jon Bratseth | 2021-07-02 | 4 | -35/+21 |
| | | |||||
* | | Add metric for render latency | Bjørn Christian Seime | 2021-07-01 | 1 | -0/+4 |
| | | |||||
* | | Rename and move 'SearchHandlerTestCase' | Bjørn Christian Seime | 2021-07-01 | 5 | -24/+22 |
|/ | |||||
* | Cast to MessageBusDocumentAccess only where needed | Jon Marius Venstad | 2021-06-30 | 1 | -0/+1 |
| | |||||
* | Use an injected VespaDocumentAccess for streaming search | Jon Marius Venstad | 2021-06-30 | 1 | -1/+1 |
| | |||||
* | Allow deviation of at least 1 document | Harald Musum | 2021-06-30 | 1 | -0/+18 |
| | | | | Let content be well-balanced when there are few docs in a cluster | ||||
* | Gc unused config options: min-group-coverage and max nodes per group | Jon Bratseth | 2021-06-15 | 1 | -1/+0 |
| | |||||
* | allow multiple tensor types for same name | Arne Juul | 2021-06-14 | 1 | -1/+25 |
| | | | | | | * with several document types you can have fields with the same name but different tensor types that all should be allowed as target for nearestNeighbor operator. | ||||
* | Let child fields override parents | Jon Bratseth | 2021-05-19 | 1 | -25/+46 |
| | |||||
* | Non-functional changes only | Jon Bratseth | 2021-05-19 | 2 | -2/+2 |
| | |||||
* | Revert "Revert "Don't consider number of working nodes in coverage"" | Jon Bratseth | 2021-05-11 | 1 | -2/+16 |
| | |||||
* | Revert "Don't consider number of working nodes in coverage" | Jon Bratseth | 2021-05-10 | 1 | -16/+2 |
| | |||||
* | Don't consider number of working nodes in coverage | Jon Bratseth | 2021-05-10 | 1 | -2/+16 |
| | | | | | | | | | | | | | | | Tring to figure out the right groups to send queries to based on the number of nodes in the group has many potential issues at times of topology changes. Since we could the number of documents available in each group by summing documents in working nodes, we do not need to also separately consider the number of working nodes in the group for correctness. Since we use adaptive dispatching by default we also do not need to consider it to avoid overloading groups with less resources available but enough documents. | ||||
* | Revert "Merge pull request #17754 from ↵ | Jon Bratseth | 2021-05-05 | 5 | -90/+62 |
| | | | | | | | vespa-engine/revert-17747-bratseth/special-tokens-take-2" This reverts commit a2c9cd4bc04f1a3eaa31524b3970b96be5c2eda9, reversing changes made to 8c61a373af0066fbdf1cca354c24b197c7347321. | ||||
* | Revert "Reapply "Bratseth/special tokens"" | Jon Bratseth | 2021-05-05 | 5 | -62/+90 |
| | |||||
* | Revert "Merge pull request #17746 from ↵ | Jon Bratseth | 2021-05-05 | 5 | -90/+62 |
| | | | | | | | vespa-engine/revert-17738-revert-17737-revert-17736-bratseth/special-tokens" This reverts commit 491856b396d003885e159345fe3f533f0fa35933, reversing changes made to 3720186303f4aef1d185525eaf61092097a64ec9. | ||||
* | Revert "Revert "Revert "Bratseth/special tokens""" | Jon Bratseth | 2021-05-05 | 5 | -62/+90 |
| | |||||
* | Revert "Revert "Bratseth/special tokens"" | Jon Bratseth | 2021-05-04 | 5 | -90/+62 |
| | |||||
* | Revert "Bratseth/special tokens" | Jon Bratseth | 2021-05-04 | 5 | -62/+90 |
| | |||||
* | Move specialtokens to linguistics | Jon Bratseth | 2021-05-04 | 4 | -37/+6 |
| | |||||
* | Remove unused file | Jon Bratseth | 2021-05-04 | 1 | -12/+0 |
| | |||||
* | Make immutable | Jon Bratseth | 2021-05-04 | 3 | -70/+85 |
| | |||||
* | Use owner's dimensions in BackedOverridableQueryProfile | Jon Bratseth | 2021-04-28 | 1 | -0/+31 |
| | | | | | | | | Use owner's dimensions rather than the backed's in BackedObverridableQueryProfile. This matters when a profile references a profile with different but overlapping dimensions and both the owner and referred profile assigns values for the same paths, within their respective dimension spaces. | ||||
* | In the case that the first item in a rank item evaluated to false it would ↵ | Henning Baldersheim | 2021-04-26 | 1 | -1/+3 |
| | | | | | | be removed, amking the next the first. That is correct behaviour for all, but the first item. If the first is false, the whole rank will be false. | ||||
* | Ensure that sameElement never accepts a WordAlternativesItem. | Henning Baldersheim | 2021-04-23 | 1 | -1/+18 |
| | |||||
* | Merge pull request #17514 from vespa-engine/bratseth/redundancy | Jon Bratseth | 2021-04-21 | 1 | -6/+6 |
|\ | | | | | Bratseth/redundancy | ||||
| * | No functional changes | Jon Bratseth | 2021-04-20 | 1 | -6/+6 |
| | | |||||
* | | Remove FS4ResourcePool | Bjørn Christian Seime | 2021-04-20 | 1 | -2/+1 |
|/ | |||||
* | Use median not average document count to determine group coverage | Jon Bratseth | 2021-04-15 | 8 | -8/+156 |
| | | | | If a group has too many nodes, all others will have less than average. | ||||
* | Change WAND to WEAKAND in debug output format | Jon Bratseth | 2021-04-14 | 5 | -13/+13 |
| | |||||
* | Collapse nesten compatible weakands | Jon Bratseth | 2021-04-13 | 1 | -0/+41 |
| | |||||
* | Merge pull request #17055 from vespa-engine/bratseth/cleanup | Henning Baldersheim | 2021-04-08 | 2 | -3/+2 |
|\ | | | | | Non-functional changes only | ||||
| * | Non-functional changes only | Jon Bratseth | 2021-03-19 | 2 | -3/+2 |
| | | |||||
* | | Handle repeated EQUIV production | Jon Bratseth | 2021-04-05 | 3 | -2/+17 |
| | | |||||
* | | Minor group logic improvements | Jon Bratseth | 2021-03-31 | 1 | -1/+1 |
| | | | | | | | | | | | | - Non-functional cleanup - Add cluster id to log messages - Don't warn that we only have N/N and require N nodes (all for the same N) | ||||
* | | Carry weight | Jon Bratseth | 2021-03-26 | 3 | -4/+4 |
| | | |||||
* | | Cleanup: No functional changes | Jon Bratseth | 2021-03-26 | 2 | -8/+9 |
| | | |||||
* | | Merge pull request #17047 from vespa-engine/bratseth/dash-in-source | Jon Bratseth | 2021-03-19 | 1 | -0/+6 |
|\ \ | | | | | | | Allow dash in ident as it is legal in source names | ||||
| * | | Allow dash in ident as it is legal in source names | Jon Bratseth | 2021-03-18 | 1 | -0/+6 |
| | | | |||||
* | | | Merge pull request #17054 from vespa-engine/bratseth/add-cast | Jon Bratseth | 2021-03-19 | 1 | -0/+6 |
|\ \ \ | | | | | | | | | Add cast to keyword list | ||||
| * | | | Add cast tokeyword list | Jon Bratseth | 2021-03-19 | 1 | -0/+6 |
| | |/ | |/| |