This PR merges the `vectors-optimize-brute-force` feature branch, which makes
the following changes to how vector functions are computed:
* Precompute the L2 norm of each vector at indexing time. (#45390)
* Switch to ByteBuffer for vector encoding. (#45936)
* Decode vectors and while computing the vector function. (#46103)
* Use an array instead of a List for the query vector. (#46155)
* Precompute the normalized query vector when using cosine similarity. (#46190)
Co-authored-by: Mayya Sharipova <mayya.sharipova@elastic.co>
This commit removes the `classic` similarity from code and docs in master (8.0). The `classic` similarity cannot be used on indices created after 7.0.
Closes#46058
This commit adds a new ShapeQueryBuilder to the xpack spatial module for
querying arbitrary Cartesian geometries indexed using the new shape field
type.
The query builder extends AbstractGeometryQueryBuilder and leverages the
ShapeQueryProcessor added in the previous field mapper commit.
Tests are provided in ShapeQueryTests in the same manner as
GeoShapeQueryTests and docs are updated to explain how the query works.
Previously, the reindex examples did not include `_doc` as the destination type.
This would result in the reindex failing with the error "Rejecting mapping
update to [users] as the final mapping would have more than 1 type: [_doc,
user]".
Relates to #43100.
This commit adds a new ShapeFieldMapper to the xpack spatial module for
indexing arbitrary cartesian geometries using a new field type called shape.
The indexing approach leverages lucene's new XYShape field type which is
backed by BKD in the same manner as LatLonShape but without the WGS84
latitude longitude restrictions. The new field mapper builds on and
extends the refactoring effort in AbstractGeometryFieldMapper and accepts
shapes in either GeoJSON or WKT format (both of which support non geospatial geometries).
Tests are provided in the ShapeFieldMapperTest class in the same manner
as GeoShapeFieldMapperTests and LegacyGeoShapeFieldMapperTests.
Documentation for how to use the new field type and what parameters are
accepted is included. The QueryBuilder for searching indexed shapes is provided in a separate commit.
Some small clarifications about force-merging and global ordinals, particularly
that global ordinals are cheap on a single-segment index and how this relates
to frozen indices.
Fixes#41687
Typically, dense vectors of both documents and queries must have the same
number of dimensions. Different number of dimensions among documents
or query vector indicate an error. This PR enforces that all vectors
for the same field have the same number of dimensions. It also enforces
that query vectors have the same number of dimensions.
A few places in the documentation had mentioned 6.7 as the version to
upgrade from, when doing an upgrade to 7.0. While this is technically
possible, this commit will replace all those mentions to 6.8, as this is
the latest version with the latest bugfixes, deprecation checks and
ugprade assistant features - which should be the one used for upgrades.
Co-Authored-By: James Rodewig <james.rodewig@elastic.co>
This commit merges the `object-fields` feature branch. The new 'flattened
object' field type allows an entire JSON object to be indexed into a field, and
provides limited search functionality over the field's contents.
It is possible for internal ML indices like `.data-frame-notifications-1` to leak,
causing other docs tests to fail when they accidentally search over these
indices. This PR updates the ignore_above tests to only search a specific index.
This PR updates the docs for `docvalue_fields` and `stored_fields` to clarify
that nested fields must be accessed through `inner_hits`. It also tweaks the
nested fields documentation to make this point more visible.
Addresses #23766.
* Previously, we mentioned multiple times that each nested object was indexed as its own document. This is repetitive, and is also a bit confusing in the context of `index.mapping.nested_fields.limit`, as that applies to the number of distinct `nested` types in the mappings, not the number of nested objects. We now just describe the issue once at the beginning of the section, to illustrate why `nested` types can be expensive.
* Reference the ongoing example to clarify the meaning of the two settings.
Addresses #28363.
The `path_match` and `path_unmatch` parameters in dynamic templates match on
object fields in addition to leaf fields. This is not obvious and can cause
surprising errors when a template is meant for a leaf field, but there are
object fields that match. This PR adds a note to the docs to describe the
current behavior.
We received some feedback that it is not completely clear why `_doc` is present
in the typeless document APIs:
> The new index APIs are PUT {index}/_doc/{id} in case of explicit ids and POST
{index}/_doc for auto-generated ids."_ Isn't this contradicting? Specifying
*types in requests is deprecated*, but we are supposed to still mention *_doc*
in write requests?
This PR updates the 'removal of types' documentation to try to clarify that
`_doc` now represents the endpoint name, as opposed to a type.
This PR makes a few clarifications to the docs for the `enabled` setting:
- Replace references to 'mapping type' with 'mapping' or 'mapping definition'.
- In code examples, clarify that the disabled fields have type `object`.
- Add a section on how disabled fields can hold non-object data.
Adds the search_as_you_type field type that acts like a text field optimized
for as-you-type search completion. It creates a couple subfields that analyze
the indexed terms as shingles, against which full terms are queried, and a
prefix subfield that analyze terms as the largest shingle size used and
edge-ngrams, against which partial terms are queried
Adds a match_bool_prefix query type that creates a boolean clause of a term
query for each term except the last, for which a boolean clause with a prefix
query is created.
The match_bool_prefix query is the recommended way of querying a search as you
type field, which will boil down to term queries for each shingle of the input
text on the appropriate shingle field, and the final (possibly partial) term
as a term query on the prefix field. This field type also supports phrase and
phrase prefix queries however
Currently if a field alias is updated, any percolator queries that contain the
alias will still refer to its old target. This PR documents the issue while we
look into addressing it.
Relates to #37212.
This adds a dedicated field mapper that supports nanosecond resolution -
at the price of a reduced date range.
When using the date field mapper, the time is stored as milliseconds since the epoch
in a long in lucene. This field mapper stores the time in nanoseconds
since the epoch - which means its range is much smaller, ranging roughly from
1970 to 2262.
Note that aggregations will still be in milliseconds.
However docvalue fields will have full nanosecond resolution
Relates #27330
There are a two major features that are not yet supported by BKD Backed geo_shape: MultiPoint queries, and CONTAINS relation. It is important we are explicitly clear in the documentation that using the new approach may not work for users that depend on these features. This commit adds an IMPORTANT NOTE section to geo_shape docs that explicitly highlights these missing features and what should be done if they are an absolute necessity.
Currently if you mix typed templates and typeless index creation or typeless
templates and typed index creation then you will end up with an error because
Elasticsearch tries to create an index that has multiple types: `_doc` and
the explicit type name that you used.
This commit proposes to give precedence to the index creation call so that
the type from the template will be ignored if the index creation call is
typeless while the template is typed, and the type from the index creation
call will be used if there is a typeless template.
This is consistent with the fact that index creation already "wins" if a field
is defined differently in the index creation call and in a template: the
definition from the index creation call is used in such cases.
Closes#37773
Ranaming as follows:
feature -> rank_feature
feature_vector -> rank_features
feature query -> rank_feature query
Ranaming is done to distinguish from other vector types.
Closes#36723
