[VEC-32] Add a few more VectorDistanceMetric (#1)

* Added index_similarity_metric option to client
* Added Jenkinsfile
* Running Jenkinsfile command in bash
* Updated version

---------

Co-authored-by: Jesse Schmidt <jschmidt@aerospike.com>

Jenkinsfile

@@ -0,0 +1,52 @@
+pipeline {
+    agent any
+
+    options {
+        disableConcurrentBuilds()
+        buildDiscarder(logRotator(numToKeepStr: '5', artifactNumToKeepStr: '5'))
+    }
+
+    stages {
+        stage("Pipeline") {
+            stages {
+                stage("Checkout") {
+                    steps {
+                        checkout([
+                                $class           : 'GitSCM',
+                                branches         : scm.branches,
+                                extensions       : scm.extensions + [[$class: 'CleanBeforeCheckout']],
+                                userRemoteConfigs: scm.userRemoteConfigs
+                        ])
+                    }
+                }
+
+                stage("Build") {
+                    steps {
+                        echo "Building.."
+                        sh "python3 -m venv .venv"
+                        sh '''#!/bin/bash
+                                source .venv/bin/activate
+                        '''
+                        sh "./proto/codegen.sh"
+                        sh "python3 -m pip install build"
+                        sh "python3 -m build"
+                    }
+                }
+
+                stage("Upload") {
+                    steps {
+                        echo "Uploading archives.."
+                        sh "python3 -m pip install twine"
+                        sh "python3 -m twine upload --repository ecosystem-python-dev-local dist/*"
+                    }
+                }
+            }
+        }
+    }
+
+    post {
+        cleanup {
+            cleanWs()
+        }
+    }
+}

proto/types.proto

@@ -29,11 +29,11 @@ message Key {
   }
 }
 
-message BoolArray {
+message BoolData {
   repeated bool value = 1;
 }
 
-message FloatArray {
+message FloatData {
   repeated float value = 1;
 }
 
@@ -70,9 +70,8 @@ message List {
 message Vector {
   // Vector data
   oneof data {
-    // TODO: rename to bool data and float data.
-    BoolArray boolArray = 1;
-    FloatArray floatArray = 2;
+    BoolData boolData = 1;
+    FloatData floatData = 2;
   }
 }
 
@@ -136,6 +135,14 @@ message IndexId {
   string namespace = 2;
 }
 
+enum VectorDistanceMetric {
+  SQUARED_EUCLIDEAN = 0;
+  COSINE = 1;
+  DOT_PRODUCT = 2;
+  MANHATTAN = 3;
+  HAMMING = 4;
+}
+
 enum IndexType {
   HNSW = 0;
 }
@@ -147,17 +154,19 @@ message IndexDefinition {
 
   IndexType type = 2;
 
+  VectorDistanceMetric vectorDistanceMetric = 3;
+
   // Optional Aerospike set name.
-  optional string set = 3;
+  optional string set = 4;
 
   // Indexed vector bin.
-  string bin = 4;
+  string bin = 5;
 
   // Number of dimensions in data.
   // Vectors not matching the dimension count will not be indexed.
-  uint32 dimensions = 5;
+  uint32 dimensions = 6;
 
-  map<string, Value> params = 6;
+  map<string, Value> params = 7;
 }
 
 // List of index definitions.

pyproject.toml

@@ -22,7 +22,7 @@ classifiers = [
     "Programming Language :: Python :: Implementation :: CPython",
     "Topic :: Database"
 ]
-version = "0.2.0"
+version = "0.3.0.dev1"
 requires-python = ">3.8"
 dependencies = [
     "grpcio",

src/aerospike_vector/conversions.py

@@ -93,9 +93,9 @@ def fromVectorDbValue(input: types_pb2.Value) -> Any:
         return [fromVectorDbValue(v) for v in input.listValue.entries]
     elif input.HasField("vectorValue"):
         vector = input.vectorValue
-        if vector.HasField("floatArray"):
-            return [v for v in vector.floatArray.value]
-        if vector.HasField("boolArray"):
-            return [v for v in vector.boolArray.value]
+        if vector.HasField("floatData"):
+            return [v for v in vector.floatData.value]
+        if vector.HasField("boolData"):
+            return [v for v in vector.boolData.value]
 
     return None

src/aerospike_vector/vectordb_admin.py

@@ -29,7 +29,9 @@ def __init__(self, seeds: types.HostPort | tuple[types.HostPort, ...],
 
     def indexCreate(self, namespace: str, name: str, set: str,
                     vector_bin_name: str, dimensions: int,
-                    params: dict[str, Any] = None):
+                    params: dict[str, Any] = None,
+                    index_similarity_metric: types_pb2.VectorDistanceMetric =
+                    types_pb2.VectorDistanceMetric.SQUARED_EUCLIDEAN):
         """Create an index"""
         index_stub = index_pb2_grpc.IndexServiceStub(
             self.channelProvider.getChannel())
@@ -47,6 +49,7 @@ def indexCreate(self, namespace: str, name: str, set: str,
             types_pb2.IndexDefinition(
                 id=types_pb2.IndexId(namespace=namespace, name=name),
                 set=set,
+                vectorDistanceMetric=index_similarity_metric,
                 bin=vector_bin_name,
                 dimensions=dimensions,
                 params=grpcParams))