From 572e6995eeab145f36e0e43a843636c77e3f8ea8 Mon Sep 17 00:00:00 2001 From: Dhruv Kaliraman Date: Mon, 24 Feb 2025 14:57:27 -0800 Subject: [PATCH] Get random hits when filtering properties --- lib/sycamore/sycamore/query/schema.py | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-) diff --git a/lib/sycamore/sycamore/query/schema.py b/lib/sycamore/sycamore/query/schema.py index 97aafd7ce..dd1405c58 100644 --- a/lib/sycamore/sycamore/query/schema.py +++ b/lib/sycamore/sycamore/query/schema.py @@ -57,7 +57,11 @@ def get_schema(self) -> OpenSearchSchema: logger.debug(f"Getting schema for index {self._index}") # Fetch example values. query["index"] = self._index - query["query"] = {"query": {"match_all": {}}, "size": self.NUM_EXAMPLES} + query["query"] = { + "query": {"match_all": {}}, + "size": self.NUM_EXAMPLES, + "sort": [{"_script": {"type": "number", "script": {"source": "Math.random()"}}}], + } random_sample = self._query_executor.query(query)["result"]["hits"]["hits"] result = OpenSearchSchema(