Merge branch 'main' of https://github.com/erikbern/ann-benchmarks

README.md

@@ -30,6 +30,7 @@ Evaluated
 * [PUFFINN](https://github.com/puffinn/puffinn) ![https://img.shields.io/github/stars/puffinn/puffinn?style=social](https://img.shields.io/github/stars/puffinn/puffinn?style=social)
 * [N2](https://github.com/kakao/n2) ![https://img.shields.io/github/stars/kakao/n2?style=social](https://img.shields.io/github/stars/kakao/n2?style=social)
 * [ScaNN](https://github.com/google-research/google-research/tree/master/scann)
+* [Vearch](https://github.com/vearch/vearch) ![https://img.shields.io/github/stars/vearch/vearch?style=social](https://img.shields.io/github/stars/vearch/vearch?style=social)
 * [Elasticsearch](https://github.com/elastic/elasticsearch) ![https://img.shields.io/github/stars/elastic/elasticsearch?style=social](https://img.shields.io/github/stars/elastic/elasticsearch?style=social): HNSW
 * [Elastiknn](https://github.com/alexklibisz/elastiknn) ![https://img.shields.io/github/stars/alexklibisz/elastiknn?style=social](https://img.shields.io/github/stars/alexklibisz/elastiknn?style=social)
 * [OpenSearch KNN](https://github.com/opensearch-project/k-NN) ![https://img.shields.io/github/stars/opensearch-project/k-NN?style=social](https://img.shields.io/github/stars/opensearch-project/k-NN?style=social)
@@ -39,9 +40,8 @@ Evaluated
 * [vald](https://github.com/vdaas/vald) ![https://img.shields.io/github/stars/vdaas/vald?style=social](https://img.shields.io/github/stars/vdaas/vald?style=social)
 * [Qdrant](https://github.com/qdrant/qdrant) ![https://img.shields.io/github/stars/qdrant/qdrant?style=social](https://img.shields.io/github/stars/qdrant/qdrant?style=social)
 * [qsgngtlib](https://gitee.com/WPJIANG-gitee/hwtl_sdu-anns-qsgngtlib)
-* [Milvus](https://github.com/milvus-io/milvus) ![https://img.shields.io/github/stars/milvus-io/milvus?style=social](https://img.shields.io/github/stars/milvus-io/milvus?style=social): [Knowhere](https://github.com/milvus-io/knowhere)
+* [Milvus](https://github.com/milvus-io/milvus) ![https://img.shields.io/github/stars/milvus-io/milvus?style=social](https://img.shields.io/github/stars/milvus-io/milvus?style=social): [Knowhere](https://github.com/milvus-io/knowhere) [Glass](https://github.com/hhy3/pyglass) 
 * [pgvector](https://github.com/pgvector/pgvector) ![https://img.shields.io/github/stars/pgvector/pgvector?style=social](https://img.shields.io/github/stars/pgvector/pgvector?style=social)
-* [Glass](https://github.com/hhy3/pyglass) 
 
 Data sets
 =========

ann_benchmarks/algorithms/elasticsearch/module.py

@@ -19,7 +19,7 @@ def __init__(self, metric: str, dimension: int, index_options: dict):
         self.num_candidates = 100
 
         index_options_str = "-".join(sorted(f"{k}-{v}" for k, v in self.index_options.items()))
-        self.name = f"es-{metric}-{dimension}-{index_options_str}"
+        self.index_name = f"{metric}-{dimension}-{index_options_str}"
         self.similarity_metric = self._vector_similarity_metric(metric)
 
         self.client = Elasticsearch(["http://localhost:9200"])
@@ -72,22 +72,22 @@ def fit(self, X):
                 },
             },
         }
-        self.client.indices.create(index=self.name, settings=settings, mappings=mappings)
+        self.client.indices.create(index=self.index_name, settings=settings, mappings=mappings)
 
         def gen():
             for i, vec in enumerate(X):
-                yield {"_op_type": "index", "_index": self.name, "id": str(i), "vec": vec.tolist()}
+                yield {"_op_type": "index", "_index": self.index_name, "id": str(i), "vec": vec.tolist()}
 
         print("Indexing ...")
         (_, errors) = bulk(self.client, gen(), chunk_size=500, request_timeout=90)
         if len(errors) != 0:
             raise RuntimeError("Failed to index documents")
 
         print("Force merge index ...")
-        self.client.indices.forcemerge(index=self.name, max_num_segments=1, request_timeout=900)
+        self.client.indices.forcemerge(index=self.index_name, max_num_segments=1, request_timeout=900)
 
         print("Refreshing index ...")
-        self.client.indices.refresh(index=self.name, request_timeout=900)
+        self.client.indices.refresh(index=self.index_name, request_timeout=900)
 
     def set_query_arguments(self, num_candidates):
         self.num_candidates = num_candidates
@@ -105,7 +105,7 @@ def query(self, q, n):
             }
         }
         res = self.client.search(
-            index=self.name,
+            index=self.index_name,
             body=body,
             size=n,
             _source=False,
@@ -121,3 +121,6 @@ def batch_query(self, X, n):
 
     def get_batch_results(self):
         return self.batch_res
+
+    def __str__(self):
+        return f"Elasticsearch(index_options: {self.index_options}, num_canditates: {self.num_candidates})"

ann_benchmarks/algorithms/opensearchknn/module.py

@@ -14,8 +14,9 @@ def __init__(self, metric, dimension, method_param):
         self.dimension = dimension
         self.method_param = method_param
         self.param_string = "-".join(k + "-" + str(v) for k, v in self.method_param.items()).lower()
-        self.name = f"os-{self.param_string}"
+        self.index_name = f"os-{self.param_string}"
         self.client = OpenSearch(["http://localhost:9200"])
+        self.ef_search = None
         self._wait_for_health_status()
 
     def _wait_for_health_status(self, wait_seconds=30, status="yellow"):
@@ -53,37 +54,38 @@ def fit(self, X):
             }
         }
 
-        self.client.indices.create(self.name, body=body)
-        self.client.indices.put_mapping(mapping, self.name)
+        self.client.indices.create(self.index_name, body=body)
+        self.client.indices.put_mapping(mapping, self.index_name)
 
-        print("Uploading data to the Index:", self.name)
+        print("Uploading data to the Index:", self.index_name)
 
         def gen():
             for i, vec in enumerate(tqdm(X)):
-                yield {"_op_type": "index", "_index": self.name, "vec": vec.tolist(), "id": str(i + 1)}
+                yield {"_op_type": "index", "_index": self.index_name, "vec": vec.tolist(), "id": str(i + 1)}
 
         (_, errors) = bulk(self.client, gen(), chunk_size=500, max_retries=2, request_timeout=10)
         assert len(errors) == 0, errors
 
         print("Force Merge...")
-        self.client.indices.forcemerge(self.name, max_num_segments=1, request_timeout=1000)
+        self.client.indices.forcemerge(self.index_name, max_num_segments=1, request_timeout=1000)
 
         print("Refreshing the Index...")
-        self.client.indices.refresh(self.name, request_timeout=1000)
+        self.client.indices.refresh(self.index_name, request_timeout=1000)
 
         print("Running Warmup API...")
-        res = urlopen(Request("http://localhost:9200/_plugins/_knn/warmup/" + self.name + "?pretty"))
+        res = urlopen(Request("http://localhost:9200/_plugins/_knn/warmup/" + self.index_name + "?pretty"))
         print(res.read().decode("utf-8"))
 
     def set_query_arguments(self, ef):
+        self.ef_search = ef
         body = {"settings": {"index": {"knn.algo_param.ef_search": ef}}}
         self.client.indices.put_settings(body=body)
 
     def query(self, q, n):
         body = {"query": {"knn": {"vec": {"vector": q.tolist(), "k": n}}}}
 
         res = self.client.search(
-            index=self.name,
+            index=self.index_name,
             body=body,
             size=n,
             _source=False,
@@ -102,4 +104,7 @@ def get_batch_results(self):
         return self.batch_res
 
     def freeIndex(self):
-        self.client.indices.delete(index=self.name)
+        self.client.indices.delete(index=self.index_name)
+
+    def __str__(self):
+        return f"OpenSearch(index_options: {self.method_param}, ef_search: {self.ef_search})"