From dd9a4c07573efb36a06d30eda837109c1890adcd Mon Sep 17 00:00:00 2001
From: Anshu Avinash <anshu.avinash35@gmail.com>
Date: Mon, 29 Jul 2024 16:42:33 +0530
Subject: [PATCH] Use efficient filtering for opensearch (#167)

Current filtering query does post-filtering which is not ideal and
resulted in low precision.

Results with post filtering:
```
{
  "params": {
    "dataset": "arxiv-titles-384-angular-filters",
    "experiment": "opensearch-default",
    "engine": "opensearch",
    "parallel": 1,
    "config": {
      "knn.algo_param.ef_search": 128
    }
  },
  "results": {
    "total_time": 708.6168032020068,
    "mean_time": 0.07045893384491791,
    "mean_precisions": 0.11399200000000001,
    "std_time": 0.06840096039381999,
    "min_time": 0.008397486002650112,
    "max_time": 3.3753458530118223,
    "rps": 14.111999538838594,
    "p95_time": 0.18164870390755816,
    "p99_time": 0.20864198897208555
  }
}
```

Results with new efficient filtering:
```
{
  "params": {
    "dataset": "arxiv-titles-384-angular-filters",
    "experiment": "opensearch-default",
    "engine": "opensearch",
    "parallel": 1,
    "config": {
      "knn.algo_param.ef_search": 128
    }
  },
  "results": {
    "total_time": 394.4290532110026,
    "mean_time": 0.03913764159695711,
    "mean_precisions": 0.610144,
    "std_time": 0.05352479065894972,
    "min_time": 0.0009066620114026591,
    "max_time": 2.1307434440095676,
    "rps": 25.35310195481576,
    "p95_time": 0.1274049270534305,
    "p99_time": 0.2078282342318563
  }
}
```
---
 engine/clients/opensearch/search.py | 7 +------
 1 file changed, 1 insertion(+), 6 deletions(-)

diff --git a/engine/clients/opensearch/search.py b/engine/clients/opensearch/search.py
index a3e36058..fc7b5cbf 100644
--- a/engine/clients/opensearch/search.py
+++ b/engine/clients/opensearch/search.py
@@ -59,12 +59,7 @@ def search_one(cls, query: Query, top: int) -> List[Tuple[int, float]]:
 
         meta_conditions = cls.parser.parse(query.meta_conditions)
         if meta_conditions:
-            opensearch_query = {
-                "bool": {
-                    "must": [opensearch_query],
-                    "filter": meta_conditions,
-                }
-            }
+            opensearch_query["knn"]["vector"]["filter"] = meta_conditions
 
         res = cls.client.search(
             index=OPENSEARCH_INDEX,