Added BasicDiskVectorDatabase to provide a basic automatically disk persistent vector database

Author: crpietschmann

CHANGELOG.md

@@ -5,6 +5,12 @@ All notable changes to this project will be documented in this file.
 The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.1.0/),
 and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
 
+## v2.2.0
+
+Add:
+
+- Added `BasicDiskVectorDatabase` to provide a basic automatically disk persistent vector database, and associated `BasicDiskVectorStore` and `BasicDiskVocabularyStore`.
+
 ## v2.1.3
 
 Add:

src/Build5Nines.SharpVector/BasicDiskVectorDatabase.cs

@@ -0,0 +1,47 @@
+using Build5Nines.SharpVector.Vocabulary;
+using Build5Nines.SharpVector.Id;
+using Build5Nines.SharpVector.Preprocessing;
+using Build5Nines.SharpVector.Vectorization;
+using Build5Nines.SharpVector.VectorCompare;
+using Build5Nines.SharpVector.VectorStore;
+
+namespace Build5Nines.SharpVector;
+
+/// <summary>
+/// A basic disk-backed vector database using Bag-of-Words, Cosine similarity,
+/// disk-backed vector store and vocabulary store. Uses int IDs and string metadata.
+/// </summary>
+public class BasicDiskVectorDatabase<TMetadata>
+     : DiskMemoryVectorDatabaseBase<
+        int,
+        TMetadata,
+        BasicDiskVectorStore<int, TMetadata, BasicDiskVocabularyStore<string>, string, int>,
+        BasicDiskVocabularyStore<string>,
+        string, int,
+        IntIdGenerator,
+        BasicTextPreprocessor,
+        BagOfWordsVectorizer<string, int>,
+        CosineSimilarityVectorComparer
+        >, IMemoryVectorDatabase<int, TMetadata>, IVectorDatabase<int, TMetadata>
+{
+    public BasicDiskVectorDatabase(string rootPath)
+        : base(
+            new BasicDiskVectorStore<int, TMetadata, BasicDiskVocabularyStore<string>, string, int>(
+                rootPath,
+                new BasicDiskVocabularyStore<string>(rootPath)
+            )
+        )
+    { }
+
+    [Obsolete("Use DeserializeFromBinaryStreamAsync instead.")]
+    public override async Task DeserializeFromJsonStreamAsync(Stream stream)
+    {
+        await DeserializeFromBinaryStreamAsync(stream);
+    }
+
+    [Obsolete("Use DeserializeFromBinaryStream instead.")]
+    public override void DeserializeFromJsonStream(Stream stream)
+    {
+        DeserializeFromBinaryStream(stream);
+    }
+}

src/Build5Nines.SharpVector/Build5Nines.SharpVector.csproj

@@ -9,7 +9,7 @@
     <PackageId>Build5Nines.SharpVector</PackageId>
     <PackageProjectUrl>https://sharpvector.build5nines.com</PackageProjectUrl>
     <RepositoryUrl>https://github.com/Build5Nines/SharpVector</RepositoryUrl>
-    <Version>2.1.3</Version>
+    <Version>2.2.0</Version>
     <Description>Lightweight In-memory Vector Database to embed in any .NET Applications</Description>
     <Copyright>Copyright (c) 2025 Build5Nines LLC</Copyright>
     <PackageReadmeFile>README.md</PackageReadmeFile>

src/Build5Nines.SharpVector/DiskMemoryVectorDatabaseBase.cs

@@ -0,0 +1,29 @@
+using Build5Nines.SharpVector.Id;
+using Build5Nines.SharpVector.Preprocessing;
+using Build5Nines.SharpVector.Vocabulary;
+using Build5Nines.SharpVector.Vectorization;
+using Build5Nines.SharpVector.VectorCompare;
+using Build5Nines.SharpVector.VectorStore;
+
+namespace Build5Nines.SharpVector;
+
+/// <summary>
+/// Base class for an on-disk vector database. Mirrors MemoryVectorDatabaseBase generic composition
+/// while using disk-backed stores for persistence.
+/// </summary>
+public abstract class DiskMemoryVectorDatabaseBase<TId, TMetadata, TVectorStore, TVocabularyStore, TVocabularyKey, TVocabularyValue, TIdGenerator, TTextPreprocessor, TVectorizer, TVectorComparer>
+    : MemoryVectorDatabaseBase<TId, TMetadata, TVectorStore, TVocabularyStore, TVocabularyKey, TVocabularyValue, TIdGenerator, TTextPreprocessor, TVectorizer, TVectorComparer>
+    where TId : notnull
+    where TVocabularyKey : notnull
+    where TVocabularyValue : notnull
+    where TVectorStore : IVectorStoreWithVocabulary<TId, TMetadata, TVocabularyStore, TVocabularyKey, TVocabularyValue>
+    where TVocabularyStore : IVocabularyStore<TVocabularyKey, TVocabularyValue>
+    where TIdGenerator : IIdGenerator<TId>, new()
+    where TTextPreprocessor : ITextPreprocessor<TVocabularyKey>, new()
+    where TVectorizer : IVectorizer<TVocabularyKey, TVocabularyValue>, new()
+    where TVectorComparer : IVectorComparer, new()
+{
+    protected DiskMemoryVectorDatabaseBase(TVectorStore vectorStore)
+        : base(vectorStore)
+    { }
+}