| Interface | Description |
|---|---|
| EvaluationMetric |
Implementations of
EvaluationMetric need to provide a way to compute the quality metric for
a result list returned by some search (@link SearchHits) and a list of rated documents. |
| MetricDetail |
Details about a specific
EvaluationMetric that should be included in the resonse. |
| Class | Description |
|---|---|
| DiscountedCumulativeGain |
Metric implementing Discounted Cumulative Gain.
|
| EvalQueryQuality |
Result of the evaluation metric calculation on one particular query alone.
|
| MeanReciprocalRank |
Metric implementing Mean Reciprocal Rank (https://en.wikipedia.org/wiki/Mean_reciprocal_rank).
By default documents with a rating equal or bigger than 1 are considered to be "relevant" for the reciprocal rank calculation. |
| MeanReciprocalRank.Detail | |
| PrecisionAtK |
Metric implementing Precision@K
(https://en.wikipedia.org/wiki/Information_retrieval#Precision_at_K).
By default documents with a rating equal or bigger than 1 are considered to be "relevant" for this calculation. |
| PrecisionAtK.Detail | |
| RankEvalAction |
Action for explaining evaluating search ranking results.
|
| RankEvalNamedXContentProvider | |
| RankEvalPlugin | |
| RankEvalRequest |
Request to perform a search ranking evaluation.
|
| RankEvalRequestBuilder | |
| RankEvalResponse |
Returns the results for a
RankEvalRequest.The repsonse contains a detailed section for each evaluation query in the request and possible failures that happened when executin individual queries. |
| RankEvalSpec |
Specification of the ranking evaluation request.
This class groups the queries to evaluate, including their document ratings, and the evaluation metric including its parameters. |
| RatedDocument |
Represents a document (specified by its _index/_id) and its corresponding rating
with respect to a specific search query.
|
| RatedRequest |
Definition of a particular query in the ranking evaluation request.
This usually represents a single user search intent and consists of an id (ideally human readable and referencing the search intent), the list of indices to be queries and the SearchSourceBuilder that will be used
to create the search request for this search intent.Alternatively, a template id and template parameters can be provided instead. Finally, a list of rated documents for this query also needs to be provided. |
| RatedSearchHit |
Combines a
SearchHit with a document rating. |
| RestRankEvalAction |
{
"requests": [{
"id": "amsterdam_query",
"request": {
"query": {
"match": {
"text": "amsterdam"
}
}
},
"ratings": [{
"_index": "foo",
"_id": "doc1",
"rating": 0
},
{
"_index": "foo",
"_id": "doc2",
"rating": 1
},
{
"_index": "foo",
"_id": "doc3",
"rating": 1
}
]
},
{
"id": "berlin_query",
"request": {
"query": {
"match": {
"text": "berlin"
}
},
"size": 10
},
"ratings": [{
"_index": "foo",
"_id": "doc1",
"rating": 1
}]
}
],
"metric": {
"precision": {
"ignore_unlabeled": true
}
}
}
|
| TransportRankEvalAction |
Instances of this class execute a collection of search intents (read: user
supplied query parameters) against a set of possible search requests (read:
search specifications, expressed as query/search request templates) and
compares the result against a set of annotated documents per search intent.
|