Adding support for the `search_type` request parameter to the Ranking Evaluation
API since this parameter can impact the ranking and the metric score and should
be choosen in the same way when evaluating the search as later in the real
search.
Closes#48503
* Convert RunTask to use testclusers, remove ClusterFormationTasks
This PR adds a new RunTask and a way for it to start a
testclusters cluster out of band and block on it to replace
the old RunTask that used ClusterFormationTasks.
With this we can now remove ClusterFormationTasks.
The base classes for transport requests and responses currently
implement Streamable and Writeable. The writeTo method on these base
classes is implemented with an empty implementation. Not only does this
complicate subclasses to think they need to call super.writeTo, but it
also can lead to not implementing writeTo when it should have been
implemented, or extendiong one of these classes when not necessary,
since there is nothing to actually implement.
This commit removes the empty writeTo from these base classes, and fixes
subclasses to not call super and in some cases implement an empty
writeTo themselves.
relates #34389
Action is a class that encapsulates meta information about an action
that allows it to be called remotely, specifically the action name and
response type. With recent refactoring, the action class can now be
constructed as a static constant, instead of needing to create a
subclass. This makes the old pattern of creating a singleton INSTANCE
both misnamed and lacking a common placement.
This commit renames Action to ActionType, thus allowing the old INSTANCE
naming pattern to be TYPE on the transport action itself. ActionType
also conveys that this class is also not the action itself, although
this change does not rename any concrete classes as those will be
removed organically as they are converted to TYPE constants.
relates #34389
The Action base class currently works for both Streamable and Writeable
response types. This commit intorduces StreamableResponseAction, for
which only the legacy Action implementions which provide newResponse()
will extend. This eliminates the need for overriding newResponse() with
an UnsupportedOperationException.
relates #34389
Currently tests still use custom type names. In preparation for the final types
removal this change moves all of them to use the default "_doc" type in tests.
This PR attempts to remove all typed calls from our YAML REST tests. The PR adds include_type_name: false to create index requests that use a mapping and also to put mapping requests. It also removes _type from index requests where they haven't already been removed. The PR ignores tests named *_with_types.yml since this are specifically testing typed API behaviour.
The change also includes changing the test harness to add the type _doc to index, update, get and bulk requests that do not specify the document type when the test is running against a mixed 7.x/6.x cluster.
* Remove empty statements
There are a couple of instances of undocumented empty statements all across the
code base. While they are mostly harmless, they make the code hard to read and
are potentially error-prone. Removing most of these instances and marking blocks
that look empty by intention as such.
* Change test, slightly more verbose but less confusing
Today if a wildcard, date-math expression or alias expands/resolves
to an index that is search-throttled we still search it. This is likely
not the desired behavior since it can unexpectedly slow down searches
significantly.
This change adds a new indices option that allows `search`, `count`
and `msearch` to ignore throttled indices by default. Users can
force expansion to throttled indices by using `ignore_throttled=true`
on the rest request to expand also to throttled indices.
Relates to #34352
With this commit we differentiate between permanent circuit breaking
exceptions (which require intervention from an operator and should not
be automatically retried) and transient ones (which may heal themselves
eventually and should be retried). Furthermore, the parent circuit
breaker will categorize a circuit breaking exception as either transient
or permanent based on the categorization of memory usage of its child
circuit breakers.
Closes#31986
Relates #34460
Optionals containing boxed primitive types are prohibitively costly because they
have two level of boxing. For Optional<Integer> the analogous OptionalInt can be
used to avoid the boxing of the contained int value.
This adds the ERR metric to the provided xContent parsers in the module and the
high level rest client registry. Also adding integration tests to make sure the
metric is correctly registered and usable from the client.
The notion of "quality" is an overloaded term in the search ranking evaluation
context. Its usually used to decribe certain levels of "good" vs. "bad" of a
seach result with respect to the users information need. We currently report the
result of the ranking evaluation as `quality_level` which is a bit missleading.
This changes the response parameter name to `metric_score` which fits better.
Currently the ranking evaluation response contains a 'unknown_docs' section
for each search use case in the evaluation set. It contains document ids for
results in the search hits that currently don't have a quality rating.
This change renames it to `unrated_docs`, which better reflects its purpose.
This change adds Expected Reciprocal Rank (ERR) as a ranking evaluation metric
as descriped in:
Chapelle, O., Metlzer, D., Zhang, Y., & Grinspan, P. (2009).
Expected reciprocal rank for graded relevance.
Proceeding of the 18th ACM Conference on Information and Knowledge Management.
https://doi.org/10.1145/1645953.1646033
ERR is an extension of the classical reciprocal rank to the graded relevance
case and assumes a cascade browsing model. It quantifies the usefulness of a
document at rank `i` conditioned on the degree of relevance of the items at ranks
less than `i`. ERR seems to be gain traction as an alternative to (n)DCG, so it
seems like a good metric to support. Also ERR seems to be the default optimization
metric used for training in RankLib, a widely used learning to rank library.
Relates to #29653
TransportAction currently contains 2 doExecute methods, one which takes
a the task, and one that does not. The latter is what some subclasses
implement, while the first one just calls the latter, dropping the given
task. This commit combines these methods, in favor of just always
assuming a task is present.
Most transport actions don't need the node ThreadPool. This commit
removes the ThreadPool as a super constructor parameter for
TransportAction. The actions that do need the thread pool then have a
member added to keep it from their own constructor.
Most transport actions don't need to resolve index names. This commit
removes the index name resolver as a super constructor parameter for
TransportAction. The actions that do need the resolver then have a
member added to keep the resolver from their own constructor.
Since #30966, Action no longer has anything but a call to the
GenericAction super constructor. This commit renames GenericAction
into Action, thus eliminating the Action class. Additionally, this
commit removes the Request generic parameter of the class, since
it was unused.
This change moves tests in `smoke-test-rank-eval-with-mustache` into the main
ranking evaluation module by declaring that the integration testing cluster
requires the `lang-mustache` plugin. This avoids having to maintain the qa
project for only one basic test suite.
While the other two ranking evaluation metrics (precicion and reciprocal rank)
already provide a more detailed output for how their score is calculated, the
discounted cumulative gain metric (dcg) and its normalized variant are lacking
this until now. Its not really clear which level of detail might be useful for
debugging and understanding the final metric calculation, but this change adds a
`metric_details` section to REST output that contains some information about the
evaluation details.
ObjectParser should throw XContentParseExceptions, not IAE. A dedicated parsing
exception can includes the place where the error occurred.
Closes#30605
This commit removes the RequestBuilder generic type from Action. It was
needed to be used by the newRequest method, which in turn was used by
client.prepareExecute. Both of these methods are now removed, along with
the existing users of prepareExecute constructing the appropriate
builder directly.
Currently the ranking evaluation API accepts the full query syntax for
the queries specified in the evaluation set and executes them via multi
search. This potentially runs costly aggregations and suggestions too.
This change adds checks that forbid using aggregations, suggesters,
highlighters and the explain and profile options in the queries that are
run as part of the ranking evaluation since they are irrelevent in the
context of this API.
This *mostly* silences `javadoc`'s warning about defaulting to
generating html4 files by enabling generating html5 file for the
projects for which that works. It didn't work in a half dozen projects,
about half of which I've fixed in this PR, entirely by replacing
`<tt>thing</tt>` with `{@code thing}`.
There are a few remaining projects that contain javadoc with invalid
html5. I'll fix those projects in a followup.
The ranking evaluation requests so far were not tested against aliases
but they should run regardless of the targeted index is a real index or
an alias. This change adds cases for this to the integration and rest
tests.
Allow high level java rest client to access details of the metric
calculation by making them accessible across packages. Also renaming the
inner `Breakdown` classes of the evaluation metrics to `Detail` to
better communicate their use.
This change validates that the `_search` request does not have trailing
tokens after the main object and fails the request with a parsing exception otherwise.
Closes#28995
* Move ObjectParser into the x-content lib
This moves `ObjectParser`, `AbstractObjectParser`, and
`ConstructingObjectParser` into the libs/x-content dependency. This decoupling
allows them to be used for parsing for projects that don't want to depend on the
entire Elasticsearch jar.
Relates to #28504
Currently the ranking evaluation API doesn't support many of the
standard parameters of the search API. Some of these make sense, like
adding support for the common indices options parameters, which this
change adds.
Fixes and edge case where DiscountedCumulativeGain can return NaN as
result of the quality metric calculation. This can happen when the
search result set is empty and normalization is used. We should return 0
in this case. Also adding related unit tests to the other two metrics.
Christoph Büscher
Alpar Torok
Jim Ferenczi
Armin Braun
Ryan Ernst
David Turner
Colin Goodheart-Smithe
Luca Cavanna
Simon Willnauer
Alexander Reelsen
Daniel Mitterdorfer
Kazuhiro Sera
Ryan Ernst
Nik Everett
Lee Hinman