446 Commits

Author SHA1 Message Date
Kishore Nallan
995fae5123 Merge branch 'v0.24.1' into v0.25 2023-02-10 11:39:11 +05:30
Kishore Nallan
eabea8f248 Handle geopoint inside nested array of objects. 2023-02-07 20:11:40 +05:30
Kishore Nallan
8d9825617b Merge branch 'v0.24.1' into v0.25 2023-02-06 10:24:05 +05:30
Kishore Nallan
86535e24aa Improve error message for nested array object string field. 2023-02-05 13:20:02 +05:30
Kishore Nallan
f40637fe64 Enabling exhaustive search should automatically drop tokens. 2023-02-04 21:27:25 +05:30
Kishore Nallan
d44afc3295 Enabling exhaustive search should automatically drop tokens. 2023-02-02 16:48:18 +05:30
Kishore Nallan
20997c9be8 Merge branch 'main' into v0.25
# Conflicts:
#	test/collection_specific_more_test.cpp
2023-01-31 07:23:27 +05:30
Kishore Nallan
f3d4bb4924 Limit text indexed for string sorting to prevent abuse. 2023-01-27 15:21:56 +05:30
Krunal Gandhi
e6c3017352
Single value facet (#854)
* single valued facet initial commit

* refactoring and bug fixes

* modularizing repeatative code to function call
2023-01-13 13:21:06 +05:30
Kishore Nallan
cd5cfc5445 Merge branch 'v0.24-nested' into v0.25
# Conflicts:
#	include/collection.h
#	src/collection.cpp
#	src/collection_manager.cpp
#	src/index.cpp
2023-01-09 16:06:34 +05:30
Kishore Nallan
ff8da56d56 Implement search cutoff for faceting op. 2023-01-05 16:20:55 +05:30
Kishore Nallan
679a6cc62f Fix k value for id based vector search. 2023-01-05 14:08:46 +05:30
Kishore Nallan
bc31be874a Add text match modes: max_score and max_weight. 2023-01-04 20:30:30 +05:30
Kishore Nallan
5c75d1553a Merge branch 'bazel-build' into v0.25 2023-01-03 11:02:09 +05:30
Kishore Nallan
b8b9fb20b3 Merge branch 'v0.24-nested' into v0.25
# Conflicts:
#	include/collection.h
#	src/collection.cpp
#	src/collection_manager.cpp
#	test/collection_faceting_test.cpp
2023-01-03 09:42:21 +05:30
Kishore Nallan
1a66a25e2f Enable slow searches logging by default (30s cutoff) 2023-01-02 18:53:36 +05:30
Kishore Nallan
bf0f7430a0 Allow vector query to pass a document ID. 2023-01-01 18:00:50 +05:30
Harpreet Sangar
34b16cdb21
Add != compatibility for numeric type. (#835)
* Add `!=` compatibility for numeric type.

* Add `numeric_not_equals_filter` method.

* Refactor `numeric_not_equals_filter` method.

* Support `!=` for numeric field multi-value filter_by.

* Ignore local settings file.

* Add comment.

* Add tests.

* Update negation test.

* Set `apply_not_equals` flag for string fields as well.

* Add `field: [!=value]` test.
2022-12-29 20:57:10 +05:30
Kishore Nallan
a10cf167ca Fix edge case for facet counts with empty strings in array. 2022-12-21 14:39:06 +05:30
Kishore Nallan
bc37cf3209 Use equal spaced sampling instead of rand. 2022-12-20 15:33:05 +05:30
Kishore Nallan
df8f6849fe Facet estimation for large result sets. 2022-12-15 16:27:01 +05:30
Krunal Gandhi
2e8ad0f7c3
Range faceting (#823)
* facet range base commit

* ammending range checks to strictly abide to range, ignoring outliers

* range facet query parsing and results

* Making facet processing generic for both
range and non-range facets

* Adding facet parsing tests and bug fixes

* Adding range facet search tests and some refactoring

* Refactoring and adding more tests for seach. adding Range should be continuous check

* Refactoring the code

* making test case asserts more verbose

Co-authored-by: Kishore Nallan <kishorenc@gmail.com>
2022-12-14 15:49:14 +05:30
Kishore Nallan
5332d77710 Fix typo/prefix regression with weights. 2022-12-07 11:32:10 +05:30
Kishore Nallan
b911766379 Implement search cutoff during intersection. 2022-11-30 15:37:12 +05:30
Kishore Nallan
44186611ae Handle vector value being null. 2022-11-24 11:51:40 +05:30
Kishore Nallan
7aa7543dae Speed up vector indexing.
Parallelize inserts + remove sort/filter indices for vector field.
2022-11-03 15:36:00 +05:30
Kishore Nallan
ae49dab8fd Vector flat search: handle missing values. 2022-10-21 16:42:23 +05:30
Kishore Nallan
4808d1610a Flat search cutoff for vector search. 2022-10-19 13:48:07 +05:30
Kishore Nallan
ff8fbc0642 Fix an issue with multi-field phrase matching. 2022-10-18 11:49:35 +05:30
Kishore Nallan
374cb6948b Reclaim deleted memory on vector insert. 2022-10-11 16:25:13 +05:30
Harpreet Sangar
f75cbca920
Filter query (#737)
* Implement advance filter query parser.

* Remove `exclude_filters`.

* Review changes

* Implement do_filtering2

* Use updated implementation

* Pass failing tests.

* Handle special characters in value.

* Add space around  in embedded filters

* Update tokenize

* Rename to `recursive_filtering`

* Deconstruct filter tree.

* Remove separating space requirement for operators

* Fix rebase errors.

* Undo filter override change

* Format changes

* Pass StaticFilteringMultipleRuleMatch

* Undo formatting changes

* Pass DynamicFilteringMultipleRuleMatch

* Review changes.

* Add filter_query guard.
2022-10-11 11:21:20 +05:30
Kishore Nallan
36a0405416 More prefix expansion tweaks. 2022-10-07 10:28:03 +05:30
Kishore Nallan
85d5ce06b2 Refactor prefix search expansion. 2022-10-05 16:14:32 +05:30
Kishore Nallan
77af30ef93 Handle prefix expansion for the same field. 2022-09-29 17:29:30 +05:30
Kishore Nallan
eaa2b5027a Remove unused arg. 2022-09-21 16:10:15 +05:30
Kishore Nallan
c8afd1e7c6 Use k value from vector query params. 2022-09-17 16:00:56 +05:30
Kishore Nallan
8895435591 Ensure that vec distance is always positive for cosine. 2022-09-16 16:48:59 +05:30
Kishore Nallan
49268d648e Persist vector field properties and restore. 2022-09-15 17:32:17 +05:30
Kishore Nallan
c7f879bf30 Return vector distance in response. 2022-09-15 11:34:27 +05:30
Kishore Nallan
31559f15b2 Numerical facet field values must be fully removed. 2022-09-14 16:22:15 +05:30
Kishore Nallan
36f38c31c1 String sorting should handle accented characters. 2022-09-11 08:40:33 +05:30
Kishore Nallan
832b519633 Optimize wildcard query sans filter/sort/facets. 2022-09-10 18:21:13 +05:30
Kishore Nallan
56344a20ca Remove spurious logs. 2022-09-08 16:38:56 +05:30
Kishore Nallan
d489702fca Handle vector index purge. 2022-09-07 10:57:52 +05:30
Kishore Nallan
a28ee865bf Fix valgrind warnings. 2022-09-03 19:23:36 +05:30
Kishore Nallan
44edeb8a5e Buffer IDs during filtering. 2022-09-03 18:35:14 +05:30
Kishore Nallan
720855f406 Use cosine similarity as default vector distance. 2022-09-02 17:48:02 +05:30
Kishore Nallan
0edee3ae12 Generify removal of field with different type due to auto schema detection. 2022-09-01 16:23:46 +05:30
Kishore Nallan
943b2ef06f Handle update of field indexed with object value, but later on as string. 2022-09-01 13:47:36 +05:30
Kishore Nallan
914a401312 Resize hnsw graph dynamically. 2022-08-27 18:52:52 +05:30