REST Resource: projects.locations.corpora.searchConfigs

Resource: SearchConfig

SearchConfig stores different properties that will affect search behaviors and search results.

JSON representation 
{
  "name": string,
  "facetProperty": {
    object (FacetProperty)
  },
  "searchCriteriaProperty": {
    object (SearchCriteriaProperty)
  }
}
Fields
name

string

Resource name of the search configuration. For CustomSearchCriteria, searchConfig would be the search operator name. For Facets, searchConfig would be the facet dimension name. Form: projects/{project_number}/locations/{location}/corpora/{corpus}/searchConfigs/{searchConfig}
facetProperty

object (FacetProperty)

Establishes a FacetDimension and associated specifications.
searchCriteriaProperty

object (SearchCriteriaProperty)

Creates a mapping between a custom SearchCriteria and one or more UGA keys.

FacetProperty

Central configuration for a facet.

JSON representation 
{
  "mappedFields": [
    string
  ],
  "displayName": string,
  "resultSize": string,
  "bucketType": enum (FacetBucketType),

  // Union field range_facet_config can be only one of the following:
  "fixedRangeBucketSpec": {
    object (FixedRangeBucketSpec)
  },
  "customRangeBucketSpec": {
    object (CustomRangeBucketSpec)
  },
  "datetimeBucketSpec": {
    object (DateTimeBucketSpec)
  }
  // End of list of possible types for union field range_facet_config.
}
Fields
mappedFields[]

string

Name of the facets, which are the dimensions users want to use to refine search results. mappedFields will match UserSpecifiedDataSchema keys.

For example, user can add a bunch of UGAs with the same key, such as player:adam, player:bob, player:charles. When multiple mappedFields are specified, will merge their value together as final facet value. E.g. home_team: a, home_team:b, away_team:a, away_team:c, when facet_field = [home_team, away_team], facet_value will be [a, b, c].

UNLESS this is a 1:1 facet dimension (mappedFields.size() == 1) AND the mapped_field equals the parent SearchConfig.name, the parent must also contain a SearchCriteriaProperty that maps to the same fields. mappedFields must not be empty.
displayName

string

Display name of the facet. To be used by UI for facet rendering.
resultSize

string (int64 format)

Maximum number of unique bucket to return for one facet. Bucket number can be large for high-cardinality facet such as "player". We only return top-n most related ones to user. If it's <= 0, the server will decide the appropriate resultSize.
bucketType

enum (FacetBucketType)

Facet bucket type e.g. value, range.

Union field range_facet_config.

range_facet_config can be only one of the following:
fixedRangeBucketSpec

object (FixedRangeBucketSpec)

Fixed range facet bucket config.
customRangeBucketSpec

object (CustomRangeBucketSpec)

Custom range facet bucket config.
datetimeBucketSpec

object (DateTimeBucketSpec)

Datetime range facet bucket config.

FixedRangeBucketSpec

If bucket type is FIXED_RANGE, specify how values are bucketized. Use FixedRangeBucketSpec when you want to create multiple buckets with equal granularities. Using integer bucket value as an example, when bucketStart = 0, bucketGranularity = 10, bucketCount = 5, this facet will be aggregated via the following buckets: [-inf, 0), [0, 10), [10, 20), [20, 30), [30, inf). Notably, bucketCount <= 1 is an invalid spec.

JSON representation 
{
  "bucketStart": {
    object (FacetValue)
  },
  "bucketGranularity": {
    object (FacetValue)
  },
  "bucketCount": integer
}
Fields
bucketStart

object (FacetValue)

Lower bound of the bucket. NOTE: Only integer type is currently supported for this field.
bucketGranularity

object (FacetValue)

Bucket granularity. NOTE: Only integer type is currently supported for this field.
bucketCount

integer

Total number of buckets.

FacetValue

Definition of a single value with generic type.

JSON representation 
{

  // Union field value can be only one of the following:
  "stringValue": string,
  "integerValue": string,
  "datetimeValue": {
    object (DateTime)
  }
  // End of list of possible types for union field value.
}
Fields

Union field value.

value can be only one of the following:
stringValue

string

String type value.
integerValue

string (int64 format)

Integer type value.
datetimeValue

object (DateTime)

Datetime type value.

CustomRangeBucketSpec

If bucket type is CUSTOM_RANGE, specify how values are bucketized. Use integer bucket value as an example, when the endpoints are 0, 10, 100, and 1000, we will generate the following facets: [-inf, 0), [0, 10), [10, 100), [100, 1000), [1000, inf). Notably: - endpoints must be listed in ascending order. Otherwise, the SearchConfig API will reject the facet config. - < 1 endpoints is an invalid spec.

JSON representation 
{
  "endpoints": [
    {
      object (FacetValue)
    }
  ]
}
Fields
endpoints[]

object (FacetValue)

Currently, only integer type is supported for this field.

DateTimeBucketSpec

If bucket type is DATE, specify how date values are bucketized.

JSON representation 
{
  "granularity": enum (Granularity)
}
Fields
granularity

enum (Granularity)

Granularity of date type facet.

Granularity

Granularity enum for the datetime bucket.

Enums
GRANULARITY_UNSPECIFIED Unspecified granularity.
YEAR Granularity is year.
MONTH Granularity is month.
DAY Granularity is day.

FacetBucketType

Different types for a facet bucket.

Enums
FACET_BUCKET_TYPE_UNSPECIFIED Unspecified type.
FACET_BUCKET_TYPE_VALUE Value type.
FACET_BUCKET_TYPE_DATETIME Datetime type.
FACET_BUCKET_TYPE_FIXED_RANGE Fixed Range type.
FACET_BUCKET_TYPE_CUSTOM_RANGE Custom Range type.

SearchCriteriaProperty

Central configuration for custom search criteria.

JSON representation 
{
  "mappedFields": [
    string
  ]
}
Fields
mappedFields[]

string

Each mapped_field corresponds to a UGA key. To understand how this property works, take the following example. In the SearchConfig table, the user adds this entry: searchConfig { name: "person" searchCriteriaProperty { mappedFields: "player" mappedFields: "coach" } }

Now, when a user issues a query like: criteria { field: "person" textArray { txtValues: "Tom Brady" txtValues: "Bill Belichick" } }

MWH search will return search documents where (player=Tom Brady || coach=Tom Brady || player=Bill Belichick || coach=Bill Belichick).

Methods

create

 Creates a search configuration inside a corpus.

delete

 Deletes a search configuration inside a corpus.

get

 Gets a search configuration inside a corpus.

list

 Lists all search configurations inside a corpus.

patch

 Updates a search configuration inside a corpus.