Something went wrong on our end. Please try again.
Created by: chwarwick
Switch aggregation search logic to using the search client instead of the api so that chunk matches are populated and can be used for capture group processing. Also moved the aggregation logic out of the stream search package into it's own.
resolves https://github.com/sourcegraph/sourcegraph/issues/40576
Updated unit tests
Manual api tests verifying counts match previous results using search API:
{
searchQueryAggregate(query: "context:global worker", patternType: lucky) {
aggregations(mode: REPO, limit: 10) {
__typename
... on ExhaustiveSearchAggregationResult {
groups {
label
count
query
}
otherResultCount
otherGroupCount
}
... on SearchAggregationNotAvailable {
reason
}
}
}
}response
{
"data": {
"searchQueryAggregate": {
"aggregations": {
"__typename": "ExhaustiveSearchAggregationResult",
"groups": [
{
"label": "github.com/sgtest/megarepo",
"count": 38500,
"query": null
},
{
"label": "github.com/sourcegraph/sourcegraph",
"count": 6604,
"query": null
},
{
"label": "github.com/sourcegraph-testing/etcd",
"count": 13,
"query": null
}
],
"otherResultCount": 0,
"otherGroupCount": 0
}
}
}
}