mirror of
https://codeberg.org/forgejo/forgejo.git
synced 2024-11-14 10:34:05 +00:00
832 lines
30 KiB
Go
832 lines
30 KiB
Go
|
// Copyright 2012-present Oliver Eilhard. All rights reserved.
|
||
|
// Use of this source code is governed by a MIT-license.
|
||
|
// See http://olivere.mit-license.org/license.txt for details.
|
||
|
|
||
|
package elastic
|
||
|
|
||
|
import (
|
||
|
"context"
|
||
|
"encoding/json"
|
||
|
"fmt"
|
||
|
"net/http"
|
||
|
"net/url"
|
||
|
"reflect"
|
||
|
"strings"
|
||
|
|
||
|
"github.com/olivere/elastic/v7/uritemplates"
|
||
|
)
|
||
|
|
||
|
// Search for documents in Elasticsearch.
|
||
|
type SearchService struct {
|
||
|
client *Client
|
||
|
|
||
|
pretty *bool // pretty format the returned JSON response
|
||
|
human *bool // return human readable values for statistics
|
||
|
errorTrace *bool // include the stack trace of returned errors
|
||
|
filterPath []string // list of filters used to reduce the response
|
||
|
headers http.Header // custom request-level HTTP headers
|
||
|
|
||
|
searchSource *SearchSource // q
|
||
|
source interface{}
|
||
|
searchType string // search_type
|
||
|
index []string
|
||
|
typ []string
|
||
|
routing string // routing
|
||
|
preference string // preference
|
||
|
requestCache *bool // request_cache
|
||
|
ignoreUnavailable *bool // ignore_unavailable
|
||
|
ignoreThrottled *bool // ignore_throttled
|
||
|
allowNoIndices *bool // allow_no_indices
|
||
|
expandWildcards string // expand_wildcards
|
||
|
lenient *bool // lenient
|
||
|
maxResponseSize int64
|
||
|
allowPartialSearchResults *bool // allow_partial_search_results
|
||
|
typedKeys *bool // typed_keys
|
||
|
seqNoPrimaryTerm *bool // seq_no_primary_term
|
||
|
batchedReduceSize *int // batched_reduce_size
|
||
|
maxConcurrentShardRequests *int // max_concurrent_shard_requests
|
||
|
preFilterShardSize *int // pre_filter_shard_size
|
||
|
restTotalHitsAsInt *bool // rest_total_hits_as_int
|
||
|
|
||
|
ccsMinimizeRoundtrips *bool // ccs_minimize_roundtrips
|
||
|
|
||
|
}
|
||
|
|
||
|
// NewSearchService creates a new service for searching in Elasticsearch.
|
||
|
func NewSearchService(client *Client) *SearchService {
|
||
|
builder := &SearchService{
|
||
|
client: client,
|
||
|
searchSource: NewSearchSource(),
|
||
|
}
|
||
|
return builder
|
||
|
}
|
||
|
|
||
|
// Pretty tells Elasticsearch whether to return a formatted JSON response.
|
||
|
func (s *SearchService) Pretty(pretty bool) *SearchService {
|
||
|
s.pretty = &pretty
|
||
|
return s
|
||
|
}
|
||
|
|
||
|
// Human specifies whether human readable values should be returned in
|
||
|
// the JSON response, e.g. "7.5mb".
|
||
|
func (s *SearchService) Human(human bool) *SearchService {
|
||
|
s.human = &human
|
||
|
return s
|
||
|
}
|
||
|
|
||
|
// ErrorTrace specifies whether to include the stack trace of returned errors.
|
||
|
func (s *SearchService) ErrorTrace(errorTrace bool) *SearchService {
|
||
|
s.errorTrace = &errorTrace
|
||
|
return s
|
||
|
}
|
||
|
|
||
|
// FilterPath specifies a list of filters used to reduce the response.
|
||
|
func (s *SearchService) FilterPath(filterPath ...string) *SearchService {
|
||
|
s.filterPath = filterPath
|
||
|
return s
|
||
|
}
|
||
|
|
||
|
// Header adds a header to the request.
|
||
|
func (s *SearchService) Header(name string, value string) *SearchService {
|
||
|
if s.headers == nil {
|
||
|
s.headers = http.Header{}
|
||
|
}
|
||
|
s.headers.Add(name, value)
|
||
|
return s
|
||
|
}
|
||
|
|
||
|
// Headers specifies the headers of the request.
|
||
|
func (s *SearchService) Headers(headers http.Header) *SearchService {
|
||
|
s.headers = headers
|
||
|
return s
|
||
|
}
|
||
|
|
||
|
// SearchSource sets the search source builder to use with this service.
|
||
|
func (s *SearchService) SearchSource(searchSource *SearchSource) *SearchService {
|
||
|
s.searchSource = searchSource
|
||
|
if s.searchSource == nil {
|
||
|
s.searchSource = NewSearchSource()
|
||
|
}
|
||
|
return s
|
||
|
}
|
||
|
|
||
|
// Source allows the user to set the request body manually without using
|
||
|
// any of the structs and interfaces in Elastic.
|
||
|
func (s *SearchService) Source(source interface{}) *SearchService {
|
||
|
s.source = source
|
||
|
return s
|
||
|
}
|
||
|
|
||
|
// Index sets the names of the indices to use for search.
|
||
|
func (s *SearchService) Index(index ...string) *SearchService {
|
||
|
s.index = append(s.index, index...)
|
||
|
return s
|
||
|
}
|
||
|
|
||
|
// Type adds search restrictions for a list of types.
|
||
|
//
|
||
|
// Deprecated: Types are in the process of being removed. Instead of using a type, prefer to
|
||
|
// filter on a field on the document.
|
||
|
func (s *SearchService) Type(typ ...string) *SearchService {
|
||
|
s.typ = append(s.typ, typ...)
|
||
|
return s
|
||
|
}
|
||
|
|
||
|
// Timeout sets the timeout to use, e.g. "1s" or "1000ms".
|
||
|
func (s *SearchService) Timeout(timeout string) *SearchService {
|
||
|
s.searchSource = s.searchSource.Timeout(timeout)
|
||
|
return s
|
||
|
}
|
||
|
|
||
|
// Profile sets the Profile API flag on the search source.
|
||
|
// When enabled, a search executed by this service will return query
|
||
|
// profiling data.
|
||
|
func (s *SearchService) Profile(profile bool) *SearchService {
|
||
|
s.searchSource = s.searchSource.Profile(profile)
|
||
|
return s
|
||
|
}
|
||
|
|
||
|
// Collapse adds field collapsing.
|
||
|
func (s *SearchService) Collapse(collapse *CollapseBuilder) *SearchService {
|
||
|
s.searchSource = s.searchSource.Collapse(collapse)
|
||
|
return s
|
||
|
}
|
||
|
|
||
|
// TimeoutInMillis sets the timeout in milliseconds.
|
||
|
func (s *SearchService) TimeoutInMillis(timeoutInMillis int) *SearchService {
|
||
|
s.searchSource = s.searchSource.TimeoutInMillis(timeoutInMillis)
|
||
|
return s
|
||
|
}
|
||
|
|
||
|
// TerminateAfter specifies the maximum number of documents to collect for
|
||
|
// each shard, upon reaching which the query execution will terminate early.
|
||
|
func (s *SearchService) TerminateAfter(terminateAfter int) *SearchService {
|
||
|
s.searchSource = s.searchSource.TerminateAfter(terminateAfter)
|
||
|
return s
|
||
|
}
|
||
|
|
||
|
// SearchType sets the search operation type. Valid values are:
|
||
|
// "dfs_query_then_fetch" and "query_then_fetch".
|
||
|
// See https://www.elastic.co/guide/en/elasticsearch/reference/7.0/search-request-search-type.html
|
||
|
// for details.
|
||
|
func (s *SearchService) SearchType(searchType string) *SearchService {
|
||
|
s.searchType = searchType
|
||
|
return s
|
||
|
}
|
||
|
|
||
|
// Routing is a list of specific routing values to control the shards
|
||
|
// the search will be executed on.
|
||
|
func (s *SearchService) Routing(routings ...string) *SearchService {
|
||
|
s.routing = strings.Join(routings, ",")
|
||
|
return s
|
||
|
}
|
||
|
|
||
|
// Preference sets the preference to execute the search. Defaults to
|
||
|
// randomize across shards ("random"). Can be set to "_local" to prefer
|
||
|
// local shards, "_primary" to execute on primary shards only,
|
||
|
// or a custom value which guarantees that the same order will be used
|
||
|
// across different requests.
|
||
|
func (s *SearchService) Preference(preference string) *SearchService {
|
||
|
s.preference = preference
|
||
|
return s
|
||
|
}
|
||
|
|
||
|
// RequestCache indicates whether the cache should be used for this
|
||
|
// request or not, defaults to index level setting.
|
||
|
func (s *SearchService) RequestCache(requestCache bool) *SearchService {
|
||
|
s.requestCache = &requestCache
|
||
|
return s
|
||
|
}
|
||
|
|
||
|
// Query sets the query to perform, e.g. MatchAllQuery.
|
||
|
func (s *SearchService) Query(query Query) *SearchService {
|
||
|
s.searchSource = s.searchSource.Query(query)
|
||
|
return s
|
||
|
}
|
||
|
|
||
|
// PostFilter will be executed after the query has been executed and
|
||
|
// only affects the search hits, not the aggregations.
|
||
|
// This filter is always executed as the last filtering mechanism.
|
||
|
func (s *SearchService) PostFilter(postFilter Query) *SearchService {
|
||
|
s.searchSource = s.searchSource.PostFilter(postFilter)
|
||
|
return s
|
||
|
}
|
||
|
|
||
|
// FetchSource indicates whether the response should contain the stored
|
||
|
// _source for every hit.
|
||
|
func (s *SearchService) FetchSource(fetchSource bool) *SearchService {
|
||
|
s.searchSource = s.searchSource.FetchSource(fetchSource)
|
||
|
return s
|
||
|
}
|
||
|
|
||
|
// FetchSourceContext indicates how the _source should be fetched.
|
||
|
func (s *SearchService) FetchSourceContext(fetchSourceContext *FetchSourceContext) *SearchService {
|
||
|
s.searchSource = s.searchSource.FetchSourceContext(fetchSourceContext)
|
||
|
return s
|
||
|
}
|
||
|
|
||
|
// Highlight adds highlighting to the search.
|
||
|
func (s *SearchService) Highlight(highlight *Highlight) *SearchService {
|
||
|
s.searchSource = s.searchSource.Highlight(highlight)
|
||
|
return s
|
||
|
}
|
||
|
|
||
|
// GlobalSuggestText defines the global text to use with all suggesters.
|
||
|
// This avoids repetition.
|
||
|
func (s *SearchService) GlobalSuggestText(globalText string) *SearchService {
|
||
|
s.searchSource = s.searchSource.GlobalSuggestText(globalText)
|
||
|
return s
|
||
|
}
|
||
|
|
||
|
// Suggester adds a suggester to the search.
|
||
|
func (s *SearchService) Suggester(suggester Suggester) *SearchService {
|
||
|
s.searchSource = s.searchSource.Suggester(suggester)
|
||
|
return s
|
||
|
}
|
||
|
|
||
|
// Aggregation adds an aggreation to perform as part of the search.
|
||
|
func (s *SearchService) Aggregation(name string, aggregation Aggregation) *SearchService {
|
||
|
s.searchSource = s.searchSource.Aggregation(name, aggregation)
|
||
|
return s
|
||
|
}
|
||
|
|
||
|
// MinScore sets the minimum score below which docs will be filtered out.
|
||
|
func (s *SearchService) MinScore(minScore float64) *SearchService {
|
||
|
s.searchSource = s.searchSource.MinScore(minScore)
|
||
|
return s
|
||
|
}
|
||
|
|
||
|
// From index to start the search from. Defaults to 0.
|
||
|
func (s *SearchService) From(from int) *SearchService {
|
||
|
s.searchSource = s.searchSource.From(from)
|
||
|
return s
|
||
|
}
|
||
|
|
||
|
// Size is the number of search hits to return. Defaults to 10.
|
||
|
func (s *SearchService) Size(size int) *SearchService {
|
||
|
s.searchSource = s.searchSource.Size(size)
|
||
|
return s
|
||
|
}
|
||
|
|
||
|
// Explain indicates whether each search hit should be returned with
|
||
|
// an explanation of the hit (ranking).
|
||
|
func (s *SearchService) Explain(explain bool) *SearchService {
|
||
|
s.searchSource = s.searchSource.Explain(explain)
|
||
|
return s
|
||
|
}
|
||
|
|
||
|
// Version indicates whether each search hit should be returned with
|
||
|
// a version associated to it.
|
||
|
func (s *SearchService) Version(version bool) *SearchService {
|
||
|
s.searchSource = s.searchSource.Version(version)
|
||
|
return s
|
||
|
}
|
||
|
|
||
|
// Sort adds a sort order.
|
||
|
func (s *SearchService) Sort(field string, ascending bool) *SearchService {
|
||
|
s.searchSource = s.searchSource.Sort(field, ascending)
|
||
|
return s
|
||
|
}
|
||
|
|
||
|
// SortWithInfo adds a sort order.
|
||
|
func (s *SearchService) SortWithInfo(info SortInfo) *SearchService {
|
||
|
s.searchSource = s.searchSource.SortWithInfo(info)
|
||
|
return s
|
||
|
}
|
||
|
|
||
|
// SortBy adds a sort order.
|
||
|
func (s *SearchService) SortBy(sorter ...Sorter) *SearchService {
|
||
|
s.searchSource = s.searchSource.SortBy(sorter...)
|
||
|
return s
|
||
|
}
|
||
|
|
||
|
// DocvalueField adds a single field to load from the field data cache
|
||
|
// and return as part of the search.
|
||
|
func (s *SearchService) DocvalueField(docvalueField string) *SearchService {
|
||
|
s.searchSource = s.searchSource.DocvalueField(docvalueField)
|
||
|
return s
|
||
|
}
|
||
|
|
||
|
// DocvalueFieldWithFormat adds a single field to load from the field data cache
|
||
|
// and return as part of the search.
|
||
|
func (s *SearchService) DocvalueFieldWithFormat(docvalueField DocvalueField) *SearchService {
|
||
|
s.searchSource = s.searchSource.DocvalueFieldWithFormat(docvalueField)
|
||
|
return s
|
||
|
}
|
||
|
|
||
|
// DocvalueFields adds one or more fields to load from the field data cache
|
||
|
// and return as part of the search.
|
||
|
func (s *SearchService) DocvalueFields(docvalueFields ...string) *SearchService {
|
||
|
s.searchSource = s.searchSource.DocvalueFields(docvalueFields...)
|
||
|
return s
|
||
|
}
|
||
|
|
||
|
// DocvalueFieldsWithFormat adds one or more fields to load from the field data cache
|
||
|
// and return as part of the search.
|
||
|
func (s *SearchService) DocvalueFieldsWithFormat(docvalueFields ...DocvalueField) *SearchService {
|
||
|
s.searchSource = s.searchSource.DocvalueFieldsWithFormat(docvalueFields...)
|
||
|
return s
|
||
|
}
|
||
|
|
||
|
// NoStoredFields indicates that no stored fields should be loaded, resulting in only
|
||
|
// id and type to be returned per field.
|
||
|
func (s *SearchService) NoStoredFields() *SearchService {
|
||
|
s.searchSource = s.searchSource.NoStoredFields()
|
||
|
return s
|
||
|
}
|
||
|
|
||
|
// StoredField adds a single field to load and return (note, must be stored) as
|
||
|
// part of the search request. If none are specified, the source of the
|
||
|
// document will be returned.
|
||
|
func (s *SearchService) StoredField(fieldName string) *SearchService {
|
||
|
s.searchSource = s.searchSource.StoredField(fieldName)
|
||
|
return s
|
||
|
}
|
||
|
|
||
|
// StoredFields sets the fields to load and return as part of the search request.
|
||
|
// If none are specified, the source of the document will be returned.
|
||
|
func (s *SearchService) StoredFields(fields ...string) *SearchService {
|
||
|
s.searchSource = s.searchSource.StoredFields(fields...)
|
||
|
return s
|
||
|
}
|
||
|
|
||
|
// TrackScores is applied when sorting and controls if scores will be
|
||
|
// tracked as well. Defaults to false.
|
||
|
func (s *SearchService) TrackScores(trackScores bool) *SearchService {
|
||
|
s.searchSource = s.searchSource.TrackScores(trackScores)
|
||
|
return s
|
||
|
}
|
||
|
|
||
|
// TrackTotalHits controls if the total hit count for the query should be tracked.
|
||
|
//
|
||
|
// See https://www.elastic.co/guide/en/elasticsearch/reference/7.1/search-request-track-total-hits.html
|
||
|
// for details.
|
||
|
func (s *SearchService) TrackTotalHits(trackTotalHits interface{}) *SearchService {
|
||
|
s.searchSource = s.searchSource.TrackTotalHits(trackTotalHits)
|
||
|
return s
|
||
|
}
|
||
|
|
||
|
// SearchAfter allows a different form of pagination by using a live cursor,
|
||
|
// using the results of the previous page to help the retrieval of the next.
|
||
|
//
|
||
|
// See https://www.elastic.co/guide/en/elasticsearch/reference/7.0/search-request-search-after.html
|
||
|
func (s *SearchService) SearchAfter(sortValues ...interface{}) *SearchService {
|
||
|
s.searchSource = s.searchSource.SearchAfter(sortValues...)
|
||
|
return s
|
||
|
}
|
||
|
|
||
|
// IgnoreUnavailable indicates whether the specified concrete indices
|
||
|
// should be ignored when unavailable (missing or closed).
|
||
|
func (s *SearchService) IgnoreUnavailable(ignoreUnavailable bool) *SearchService {
|
||
|
s.ignoreUnavailable = &ignoreUnavailable
|
||
|
return s
|
||
|
}
|
||
|
|
||
|
// IgnoreThrottled indicates whether specified concrete, expanded or aliased
|
||
|
// indices should be ignored when throttled.
|
||
|
func (s *SearchService) IgnoreThrottled(ignoreThrottled bool) *SearchService {
|
||
|
s.ignoreThrottled = &ignoreThrottled
|
||
|
return s
|
||
|
}
|
||
|
|
||
|
// AllowNoIndices indicates whether to ignore if a wildcard indices
|
||
|
// expression resolves into no concrete indices. (This includes `_all` string
|
||
|
// or when no indices have been specified).
|
||
|
func (s *SearchService) AllowNoIndices(allowNoIndices bool) *SearchService {
|
||
|
s.allowNoIndices = &allowNoIndices
|
||
|
return s
|
||
|
}
|
||
|
|
||
|
// ExpandWildcards indicates whether to expand wildcard expression to
|
||
|
// concrete indices that are open, closed or both.
|
||
|
func (s *SearchService) ExpandWildcards(expandWildcards string) *SearchService {
|
||
|
s.expandWildcards = expandWildcards
|
||
|
return s
|
||
|
}
|
||
|
|
||
|
// Lenient specifies whether format-based query failures (such as providing
|
||
|
// text to a numeric field) should be ignored.
|
||
|
func (s *SearchService) Lenient(lenient bool) *SearchService {
|
||
|
s.lenient = &lenient
|
||
|
return s
|
||
|
}
|
||
|
|
||
|
// MaxResponseSize sets an upper limit on the response body size that we accept,
|
||
|
// to guard against OOM situations.
|
||
|
func (s *SearchService) MaxResponseSize(maxResponseSize int64) *SearchService {
|
||
|
s.maxResponseSize = maxResponseSize
|
||
|
return s
|
||
|
}
|
||
|
|
||
|
// AllowPartialSearchResults indicates if an error should be returned if
|
||
|
// there is a partial search failure or timeout.
|
||
|
func (s *SearchService) AllowPartialSearchResults(enabled bool) *SearchService {
|
||
|
s.allowPartialSearchResults = &enabled
|
||
|
return s
|
||
|
}
|
||
|
|
||
|
// TypedKeys specifies whether aggregation and suggester names should be
|
||
|
// prefixed by their respective types in the response.
|
||
|
func (s *SearchService) TypedKeys(enabled bool) *SearchService {
|
||
|
s.typedKeys = &enabled
|
||
|
return s
|
||
|
}
|
||
|
|
||
|
// SeqNoPrimaryTerm specifies whether to return sequence number and
|
||
|
// primary term of the last modification of each hit.
|
||
|
func (s *SearchService) SeqNoPrimaryTerm(enabled bool) *SearchService {
|
||
|
s.seqNoPrimaryTerm = &enabled
|
||
|
return s
|
||
|
}
|
||
|
|
||
|
// BatchedReduceSize specifies the number of shard results that should be reduced
|
||
|
// at once on the coordinating node. This value should be used as a protection
|
||
|
// mechanism to reduce the memory overhead per search request if the potential
|
||
|
// number of shards in the request can be large.
|
||
|
func (s *SearchService) BatchedReduceSize(size int) *SearchService {
|
||
|
s.batchedReduceSize = &size
|
||
|
return s
|
||
|
}
|
||
|
|
||
|
// MaxConcurrentShardRequests specifies the number of concurrent shard requests
|
||
|
// this search executes concurrently. This value should be used to limit the
|
||
|
// impact of the search on the cluster in order to limit the number of
|
||
|
// concurrent shard requests.
|
||
|
func (s *SearchService) MaxConcurrentShardRequests(max int) *SearchService {
|
||
|
s.maxConcurrentShardRequests = &max
|
||
|
return s
|
||
|
}
|
||
|
|
||
|
// PreFilterShardSize specifies a threshold that enforces a pre-filter roundtrip
|
||
|
// to prefilter search shards based on query rewriting if the number of shards
|
||
|
// the search request expands to exceeds the threshold. This filter roundtrip
|
||
|
// can limit the number of shards significantly if for instance a shard can
|
||
|
// not match any documents based on it's rewrite method i.e. if date filters are
|
||
|
// mandatory to match but the shard bounds and the query are disjoint.
|
||
|
func (s *SearchService) PreFilterShardSize(threshold int) *SearchService {
|
||
|
s.preFilterShardSize = &threshold
|
||
|
return s
|
||
|
}
|
||
|
|
||
|
// RestTotalHitsAsInt indicates whether hits.total should be rendered as an
|
||
|
// integer or an object in the rest search response.
|
||
|
func (s *SearchService) RestTotalHitsAsInt(enabled bool) *SearchService {
|
||
|
s.restTotalHitsAsInt = &enabled
|
||
|
return s
|
||
|
}
|
||
|
|
||
|
// CCSMinimizeRoundtrips indicates whether network round-trips should be minimized
|
||
|
// as part of cross-cluster search requests execution.
|
||
|
func (s *SearchService) CCSMinimizeRoundtrips(enabled bool) *SearchService {
|
||
|
s.ccsMinimizeRoundtrips = &enabled
|
||
|
return s
|
||
|
}
|
||
|
|
||
|
// buildURL builds the URL for the operation.
|
||
|
func (s *SearchService) buildURL() (string, url.Values, error) {
|
||
|
var err error
|
||
|
var path string
|
||
|
|
||
|
if len(s.index) > 0 && len(s.typ) > 0 {
|
||
|
path, err = uritemplates.Expand("/{index}/{type}/_search", map[string]string{
|
||
|
"index": strings.Join(s.index, ","),
|
||
|
"type": strings.Join(s.typ, ","),
|
||
|
})
|
||
|
} else if len(s.index) > 0 {
|
||
|
path, err = uritemplates.Expand("/{index}/_search", map[string]string{
|
||
|
"index": strings.Join(s.index, ","),
|
||
|
})
|
||
|
} else if len(s.typ) > 0 {
|
||
|
path, err = uritemplates.Expand("/_all/{type}/_search", map[string]string{
|
||
|
"type": strings.Join(s.typ, ","),
|
||
|
})
|
||
|
} else {
|
||
|
path = "/_search"
|
||
|
}
|
||
|
if err != nil {
|
||
|
return "", url.Values{}, err
|
||
|
}
|
||
|
|
||
|
// Add query string parameters
|
||
|
params := url.Values{}
|
||
|
if v := s.pretty; v != nil {
|
||
|
params.Set("pretty", fmt.Sprint(*v))
|
||
|
}
|
||
|
if v := s.human; v != nil {
|
||
|
params.Set("human", fmt.Sprint(*v))
|
||
|
}
|
||
|
if v := s.errorTrace; v != nil {
|
||
|
params.Set("error_trace", fmt.Sprint(*v))
|
||
|
}
|
||
|
if len(s.filterPath) > 0 {
|
||
|
params.Set("filter_path", strings.Join(s.filterPath, ","))
|
||
|
}
|
||
|
if s.searchType != "" {
|
||
|
params.Set("search_type", s.searchType)
|
||
|
}
|
||
|
if s.routing != "" {
|
||
|
params.Set("routing", s.routing)
|
||
|
}
|
||
|
if s.preference != "" {
|
||
|
params.Set("preference", s.preference)
|
||
|
}
|
||
|
if v := s.requestCache; v != nil {
|
||
|
params.Set("request_cache", fmt.Sprint(*v))
|
||
|
}
|
||
|
if v := s.allowNoIndices; v != nil {
|
||
|
params.Set("allow_no_indices", fmt.Sprint(*v))
|
||
|
}
|
||
|
if s.expandWildcards != "" {
|
||
|
params.Set("expand_wildcards", s.expandWildcards)
|
||
|
}
|
||
|
if v := s.lenient; v != nil {
|
||
|
params.Set("lenient", fmt.Sprint(*v))
|
||
|
}
|
||
|
if v := s.ignoreUnavailable; v != nil {
|
||
|
params.Set("ignore_unavailable", fmt.Sprint(*v))
|
||
|
}
|
||
|
if v := s.ignoreThrottled; v != nil {
|
||
|
params.Set("ignore_throttled", fmt.Sprint(*v))
|
||
|
}
|
||
|
if s.seqNoPrimaryTerm != nil {
|
||
|
params.Set("seq_no_primary_term", fmt.Sprint(*s.seqNoPrimaryTerm))
|
||
|
}
|
||
|
if v := s.allowPartialSearchResults; v != nil {
|
||
|
params.Set("allow_partial_search_results", fmt.Sprint(*v))
|
||
|
}
|
||
|
if v := s.typedKeys; v != nil {
|
||
|
params.Set("typed_keys", fmt.Sprint(*v))
|
||
|
}
|
||
|
if v := s.batchedReduceSize; v != nil {
|
||
|
params.Set("batched_reduce_size", fmt.Sprint(*v))
|
||
|
}
|
||
|
if v := s.maxConcurrentShardRequests; v != nil {
|
||
|
params.Set("max_concurrent_shard_requests", fmt.Sprint(*v))
|
||
|
}
|
||
|
if v := s.preFilterShardSize; v != nil {
|
||
|
params.Set("pre_filter_shard_size", fmt.Sprint(*v))
|
||
|
}
|
||
|
if v := s.restTotalHitsAsInt; v != nil {
|
||
|
params.Set("rest_total_hits_as_int", fmt.Sprint(*v))
|
||
|
}
|
||
|
if v := s.ccsMinimizeRoundtrips; v != nil {
|
||
|
params.Set("ccs_minimize_roundtrips", fmt.Sprint(*v))
|
||
|
}
|
||
|
return path, params, nil
|
||
|
}
|
||
|
|
||
|
// Validate checks if the operation is valid.
|
||
|
func (s *SearchService) Validate() error {
|
||
|
return nil
|
||
|
}
|
||
|
|
||
|
// Do executes the search and returns a SearchResult.
|
||
|
func (s *SearchService) Do(ctx context.Context) (*SearchResult, error) {
|
||
|
// Check pre-conditions
|
||
|
if err := s.Validate(); err != nil {
|
||
|
return nil, err
|
||
|
}
|
||
|
|
||
|
// Get URL for request
|
||
|
path, params, err := s.buildURL()
|
||
|
if err != nil {
|
||
|
return nil, err
|
||
|
}
|
||
|
|
||
|
// Perform request
|
||
|
var body interface{}
|
||
|
if s.source != nil {
|
||
|
body = s.source
|
||
|
} else {
|
||
|
src, err := s.searchSource.Source()
|
||
|
if err != nil {
|
||
|
return nil, err
|
||
|
}
|
||
|
body = src
|
||
|
}
|
||
|
res, err := s.client.PerformRequest(ctx, PerformRequestOptions{
|
||
|
Method: "POST",
|
||
|
Path: path,
|
||
|
Params: params,
|
||
|
Body: body,
|
||
|
Headers: s.headers,
|
||
|
MaxResponseSize: s.maxResponseSize,
|
||
|
})
|
||
|
if err != nil {
|
||
|
return nil, err
|
||
|
}
|
||
|
|
||
|
// Return search results
|
||
|
ret := new(SearchResult)
|
||
|
if err := s.client.decoder.Decode(res.Body, ret); err != nil {
|
||
|
ret.Header = res.Header
|
||
|
return nil, err
|
||
|
}
|
||
|
ret.Header = res.Header
|
||
|
return ret, nil
|
||
|
}
|
||
|
|
||
|
// SearchResult is the result of a search in Elasticsearch.
|
||
|
type SearchResult struct {
|
||
|
Header http.Header `json:"-"`
|
||
|
TookInMillis int64 `json:"took,omitempty"` // search time in milliseconds
|
||
|
TerminatedEarly bool `json:"terminated_early,omitempty"` // request terminated early
|
||
|
NumReducePhases int `json:"num_reduce_phases,omitempty"`
|
||
|
Clusters []*SearchResultCluster `json:"_clusters,omitempty"` // 6.1.0+
|
||
|
ScrollId string `json:"_scroll_id,omitempty"` // only used with Scroll and Scan operations
|
||
|
Hits *SearchHits `json:"hits,omitempty"` // the actual search hits
|
||
|
Suggest SearchSuggest `json:"suggest,omitempty"` // results from suggesters
|
||
|
Aggregations Aggregations `json:"aggregations,omitempty"` // results from aggregations
|
||
|
TimedOut bool `json:"timed_out,omitempty"` // true if the search timed out
|
||
|
Error *ErrorDetails `json:"error,omitempty"` // only used in MultiGet
|
||
|
Profile *SearchProfile `json:"profile,omitempty"` // profiling results, if optional Profile API was active for this search
|
||
|
Shards *ShardsInfo `json:"_shards,omitempty"` // shard information
|
||
|
Status int `json:"status,omitempty"` // used in MultiSearch
|
||
|
}
|
||
|
|
||
|
// SearchResultCluster holds information about a search response
|
||
|
// from a cluster.
|
||
|
type SearchResultCluster struct {
|
||
|
Successful int `json:"successful,omitempty"`
|
||
|
Total int `json:"total,omitempty"`
|
||
|
Skipped int `json:"skipped,omitempty"`
|
||
|
}
|
||
|
|
||
|
// TotalHits is a convenience function to return the number of hits for
|
||
|
// a search result. The return value might not be accurate, unless
|
||
|
// track_total_hits parameter has set to true.
|
||
|
func (r *SearchResult) TotalHits() int64 {
|
||
|
if r.Hits != nil && r.Hits.TotalHits != nil {
|
||
|
return r.Hits.TotalHits.Value
|
||
|
}
|
||
|
return 0
|
||
|
}
|
||
|
|
||
|
// Each is a utility function to iterate over all hits. It saves you from
|
||
|
// checking for nil values. Notice that Each will ignore errors in
|
||
|
// serializing JSON and hits with empty/nil _source will get an empty
|
||
|
// value
|
||
|
func (r *SearchResult) Each(typ reflect.Type) []interface{} {
|
||
|
if r.Hits == nil || r.Hits.Hits == nil || len(r.Hits.Hits) == 0 {
|
||
|
return nil
|
||
|
}
|
||
|
var slice []interface{}
|
||
|
for _, hit := range r.Hits.Hits {
|
||
|
v := reflect.New(typ).Elem()
|
||
|
if hit.Source == nil {
|
||
|
slice = append(slice, v.Interface())
|
||
|
continue
|
||
|
}
|
||
|
if err := json.Unmarshal(hit.Source, v.Addr().Interface()); err == nil {
|
||
|
slice = append(slice, v.Interface())
|
||
|
}
|
||
|
}
|
||
|
return slice
|
||
|
}
|
||
|
|
||
|
// SearchHits specifies the list of search hits.
|
||
|
type SearchHits struct {
|
||
|
TotalHits *TotalHits `json:"total,omitempty"` // total number of hits found
|
||
|
MaxScore *float64 `json:"max_score,omitempty"` // maximum score of all hits
|
||
|
Hits []*SearchHit `json:"hits,omitempty"` // the actual hits returned
|
||
|
}
|
||
|
|
||
|
// NestedHit is a nested innerhit
|
||
|
type NestedHit struct {
|
||
|
Field string `json:"field"`
|
||
|
Offset int `json:"offset,omitempty"`
|
||
|
Child *NestedHit `json:"_nested,omitempty"`
|
||
|
}
|
||
|
|
||
|
// TotalHits specifies total number of hits and its relation
|
||
|
type TotalHits struct {
|
||
|
Value int64 `json:"value"` // value of the total hit count
|
||
|
Relation string `json:"relation"` // how the value should be interpreted: accurate ("eq") or a lower bound ("gte")
|
||
|
}
|
||
|
|
||
|
// SearchHit is a single hit.
|
||
|
type SearchHit struct {
|
||
|
Score *float64 `json:"_score,omitempty"` // computed score
|
||
|
Index string `json:"_index,omitempty"` // index name
|
||
|
Type string `json:"_type,omitempty"` // type meta field
|
||
|
Id string `json:"_id,omitempty"` // external or internal
|
||
|
Uid string `json:"_uid,omitempty"` // uid meta field (see MapperService.java for all meta fields)
|
||
|
Routing string `json:"_routing,omitempty"` // routing meta field
|
||
|
Parent string `json:"_parent,omitempty"` // parent meta field
|
||
|
Version *int64 `json:"_version,omitempty"` // version number, when Version is set to true in SearchService
|
||
|
SeqNo *int64 `json:"_seq_no"`
|
||
|
PrimaryTerm *int64 `json:"_primary_term"`
|
||
|
Sort []interface{} `json:"sort,omitempty"` // sort information
|
||
|
Highlight SearchHitHighlight `json:"highlight,omitempty"` // highlighter information
|
||
|
Source json.RawMessage `json:"_source,omitempty"` // stored document source
|
||
|
Fields map[string]interface{} `json:"fields,omitempty"` // returned (stored) fields
|
||
|
Explanation *SearchExplanation `json:"_explanation,omitempty"` // explains how the score was computed
|
||
|
MatchedQueries []string `json:"matched_queries,omitempty"` // matched queries
|
||
|
InnerHits map[string]*SearchHitInnerHits `json:"inner_hits,omitempty"` // inner hits with ES >= 1.5.0
|
||
|
Nested *NestedHit `json:"_nested,omitempty"` // for nested inner hits
|
||
|
Shard string `json:"_shard,omitempty"` // used e.g. in Search Explain
|
||
|
Node string `json:"_node,omitempty"` // used e.g. in Search Explain
|
||
|
|
||
|
// HighlightFields
|
||
|
// SortValues
|
||
|
// MatchedFilters
|
||
|
}
|
||
|
|
||
|
// SearchHitInnerHits is used for inner hits.
|
||
|
type SearchHitInnerHits struct {
|
||
|
Hits *SearchHits `json:"hits,omitempty"`
|
||
|
}
|
||
|
|
||
|
// SearchExplanation explains how the score for a hit was computed.
|
||
|
// See https://www.elastic.co/guide/en/elasticsearch/reference/7.0/search-request-explain.html.
|
||
|
type SearchExplanation struct {
|
||
|
Value float64 `json:"value"` // e.g. 1.0
|
||
|
Description string `json:"description"` // e.g. "boost" or "ConstantScore(*:*), product of:"
|
||
|
Details []SearchExplanation `json:"details,omitempty"` // recursive details
|
||
|
}
|
||
|
|
||
|
// Suggest
|
||
|
|
||
|
// SearchSuggest is a map of suggestions.
|
||
|
// See https://www.elastic.co/guide/en/elasticsearch/reference/7.0/search-suggesters.html.
|
||
|
type SearchSuggest map[string][]SearchSuggestion
|
||
|
|
||
|
// SearchSuggestion is a single search suggestion.
|
||
|
// See https://www.elastic.co/guide/en/elasticsearch/reference/7.0/search-suggesters.html.
|
||
|
type SearchSuggestion struct {
|
||
|
Text string `json:"text"`
|
||
|
Offset int `json:"offset"`
|
||
|
Length int `json:"length"`
|
||
|
Options []SearchSuggestionOption `json:"options"`
|
||
|
}
|
||
|
|
||
|
// SearchSuggestionOption is an option of a SearchSuggestion.
|
||
|
// See https://www.elastic.co/guide/en/elasticsearch/reference/7.0/search-suggesters.html.
|
||
|
type SearchSuggestionOption struct {
|
||
|
Text string `json:"text"`
|
||
|
Index string `json:"_index"`
|
||
|
Type string `json:"_type"`
|
||
|
Id string `json:"_id"`
|
||
|
Score float64 `json:"score"` // term and phrase suggesters uses "score" as of 6.2.4
|
||
|
ScoreUnderscore float64 `json:"_score"` // completion and context suggesters uses "_score" as of 6.2.4
|
||
|
Highlighted string `json:"highlighted"`
|
||
|
CollateMatch bool `json:"collate_match"`
|
||
|
Freq int `json:"freq"` // from TermSuggestion.Option in Java API
|
||
|
Source json.RawMessage `json:"_source"`
|
||
|
Contexts map[string][]string `json:"contexts,omitempty"`
|
||
|
}
|
||
|
|
||
|
// SearchProfile is a list of shard profiling data collected during
|
||
|
// query execution in the "profile" section of a SearchResult
|
||
|
type SearchProfile struct {
|
||
|
Shards []SearchProfileShardResult `json:"shards"`
|
||
|
}
|
||
|
|
||
|
// SearchProfileShardResult returns the profiling data for a single shard
|
||
|
// accessed during the search query or aggregation.
|
||
|
type SearchProfileShardResult struct {
|
||
|
ID string `json:"id"`
|
||
|
Searches []QueryProfileShardResult `json:"searches"`
|
||
|
Aggregations []ProfileResult `json:"aggregations"`
|
||
|
}
|
||
|
|
||
|
// QueryProfileShardResult is a container class to hold the profile results
|
||
|
// for a single shard in the request. It comtains a list of query profiles,
|
||
|
// a collector tree and a total rewrite tree.
|
||
|
type QueryProfileShardResult struct {
|
||
|
Query []ProfileResult `json:"query,omitempty"`
|
||
|
RewriteTime int64 `json:"rewrite_time,omitempty"`
|
||
|
Collector []interface{} `json:"collector,omitempty"`
|
||
|
}
|
||
|
|
||
|
// CollectorResult holds the profile timings of the collectors used in the
|
||
|
// search. Children's CollectorResults may be embedded inside of a parent
|
||
|
// CollectorResult.
|
||
|
type CollectorResult struct {
|
||
|
Name string `json:"name,omitempty"`
|
||
|
Reason string `json:"reason,omitempty"`
|
||
|
Time string `json:"time,omitempty"`
|
||
|
TimeNanos int64 `json:"time_in_nanos,omitempty"`
|
||
|
Children []CollectorResult `json:"children,omitempty"`
|
||
|
}
|
||
|
|
||
|
// ProfileResult is the internal representation of a profiled query,
|
||
|
// corresponding to a single node in the query tree.
|
||
|
type ProfileResult struct {
|
||
|
Type string `json:"type"`
|
||
|
Description string `json:"description,omitempty"`
|
||
|
NodeTime string `json:"time,omitempty"`
|
||
|
NodeTimeNanos int64 `json:"time_in_nanos,omitempty"`
|
||
|
Breakdown map[string]int64 `json:"breakdown,omitempty"`
|
||
|
Children []ProfileResult `json:"children,omitempty"`
|
||
|
}
|
||
|
|
||
|
// Aggregations (see search_aggs.go)
|
||
|
|
||
|
// Highlighting
|
||
|
|
||
|
// SearchHitHighlight is the highlight information of a search hit.
|
||
|
// See https://www.elastic.co/guide/en/elasticsearch/reference/7.0/search-request-highlighting.html
|
||
|
// for a general discussion of highlighting.
|
||
|
type SearchHitHighlight map[string][]string
|