BLOOM now uses a unified search service rooted in TapDB-backed access patterns.
Search v2 supports mixed record types in one query:
instancetemplatelineageaudit
It uses ILIKE + JSON-path filters in phase 1 (no FTS/trigram migrations).
POST /api/v1/search/v2/query
Body (SearchRequest):
{
"query": "sample",
"record_types": ["instance", "template"],
"categories": ["file"],
"page": 1,
"page_size": 50,
"sort_by": "timestamp",
"sort_order": "desc"
}POST /api/v1/search/v2/export
Body (SearchExportRequest):
{
"search": {
"query": "sample",
"record_types": ["instance"],
"page": 1,
"page_size": 100
},
"format": "json",
"include_metadata": true,
"max_export_rows": 10000
}Formats:
jsontsv
Legacy endpoints are still available for compatibility and include deprecation headers:
GET /api/v1/search/GET /api/v1/search/export
Successor endpoint:
POST /api/v1/search/v2/query
/search is now backed by search v2 and shows unified/faceted results.
Dewey metadata search forms now post to /search and are translated into search v2 filter requests.
Legacy Dewey endpoints are retired:
POST /search_files->410 GonePOST /search_file_sets->410 Gone
Use /search (GUI) or /api/v1/search/v2/query (API) instead.