Allowed moving local data sources without embedding them again (#1091)

Co-authored-by: Thorsten Sommer <SommerEngineering@users.noreply.github.com>

app/MindWork AI Studio/Tools/Databases/IndexStore/IndexStoreClient.cs

@@ -29,6 +29,11 @@ public abstract partial class IndexStoreClient(string name, string path) : Datab
 
     public abstract Task UpsertFileAsync(string dataSourceId, EmbeddingStateFile file, CancellationToken token);
 
+    /// <summary>
+    /// Updates the location and fingerprint of a document, preserving its chunks and their ids.
+    /// </summary>
+    public abstract Task RelocateFileAsync(string dataSourceId, string parentFileId, string absolutePath, string relativePath, string fingerprint, CancellationToken token);
+
     public abstract Task DeleteFileAsync(string dataSourceId, string filePath, CancellationToken token);
 
     public abstract Task UpsertPermanentFailureAsync(string dataSourceId, PermanentIndexingFailure failure, CancellationToken token);

app/MindWork AI Studio/Tools/Databases/IndexStore/NoIndexStoreClient.cs

@@ -52,6 +52,9 @@ public sealed partial class NoIndexStoreClient(string name, string? unavailableR
 
     public override Task UpsertFileAsync(string dataSourceId, EmbeddingStateFile file, CancellationToken token) => Task.CompletedTask;
 
+    public override Task RelocateFileAsync(string dataSourceId, string parentFileId, string absolutePath, string relativePath, string fingerprint, CancellationToken token) =>
+        Task.FromException(new InvalidOperationException(unavailableReason ?? "The index of the data sources is not available."));
+
     public override Task DeleteFileAsync(string dataSourceId, string filePath, CancellationToken token) => Task.CompletedTask;
 
     public override Task UpsertPermanentFailureAsync(string dataSourceId, PermanentIndexingFailure failure, CancellationToken token) => Task.CompletedTask;

app/MindWork AI Studio/Tools/Databases/IndexStore/SqliteIndexStoreClientImplementation.cs

@@ -149,7 +149,9 @@ public sealed partial class SqliteIndexStoreClientImplementation(string name, st
                 file.LastWriteUtc,
                 file.EmbeddedAtUtc,
                 file.ChunkCount,
-                file.TranscriptKey);
+                file.TranscriptKey,
+                file.RelativePath,
+                file.ParentFileId);
         }
 
         var permanentFailures = await context.PermanentIndexingFailures
@@ -233,6 +235,18 @@ public sealed partial class SqliteIndexStoreClientImplementation(string name, st
         await context.SaveChangesAsync(token);
     }
 
+    public override async Task RelocateFileAsync(string dataSourceId, string parentFileId, string absolutePath, string relativePath, string fingerprint, CancellationToken token)
+    {
+        await using var context = this.CreateContext();
+        var file = await context.EmbeddedFiles.SingleAsync(file => file.DataSourceId == dataSourceId && file.ParentFileId == parentFileId, token);
+        // Keep the parent and chunk ids: moving a document does not change its stored vectors.
+        file.AbsolutePath = absolutePath;
+        file.RelativePath = relativePath;
+        file.FileName = Path.GetFileName(absolutePath);
+        file.Fingerprint = fingerprint;
+        await context.SaveChangesAsync(token);
+    }
+
     public override async Task DeleteFileAsync(string dataSourceId, string filePath, CancellationToken token)
     {
         await using var context = this.CreateContext();
@@ -762,4 +776,4 @@ public sealed partial class SqliteIndexStoreClientImplementation(string name, st
         client.SetLogger(databaseClientLogger);
         return client;
     }
-}
+}

app/MindWork AI Studio/Tools/Databases/VectorStore/NoVectorStoreClient.cs

@@ -37,7 +37,10 @@ public sealed class NoVectorStoreClient(string name, string? unavailableReason,
     public override Task<IReadOnlyList<VectorSearchResult>> SearchEmbeddingAsync(string storeName, IReadOnlyList<float> vector, int maxMatches, VectorSearchFilter filter, CancellationToken token) =>
         Task.FromException<IReadOnlyList<VectorSearchResult>>(this.CreateUnavailableException());
 
-    public override Task DeleteEmbeddingByFile(string storeName, string filePath, CancellationToken token) =>
+    public override Task DeleteEmbeddingByFile(string storeName, string filePath, string parentFileId, CancellationToken token) =>
+        Task.FromException(this.CreateUnavailableException());
+
+    public override Task RelocateEmbeddingFileAsync(string storeName, string parentFileId, string absolutePath, string relativePath, string fingerprint, CancellationToken token) =>
         Task.FromException(this.CreateUnavailableException());
 
     public override Task OptimizeVectorStore(string storeName, CancellationToken token) =>

app/MindWork AI Studio/Tools/Databases/VectorStore/QdrantEdgeClientImplementation.cs

@@ -27,6 +27,7 @@ public sealed class QdrantEdgeClientImplementation(
     private const string INSERT_PATH = "/system/qdrant-edge/insert";
     private const string SEARCH_PATH = "/system/qdrant-edge/search";
     private const string DELETE_FILE_PATH = "/system/qdrant-edge/delete-file";
+    private const string RELOCATE_FILE_PATH = "/system/qdrant-edge/relocate-file";
     private const string OPTIMIZE_PATH = "/system/qdrant-edge/optimize";
     private const string DELETE_STORE_PATH = "/system/qdrant-edge/delete-store";
     private const string PREPARE_BACKUP_PATH = "/system/qdrant-edge/backup/prepare";
@@ -156,8 +157,11 @@ public sealed class QdrantEdgeClientImplementation(
             token) ?? [];
     }
 
-    public override Task DeleteEmbeddingByFile(string storeName, string filePath, CancellationToken token) =>
-        rustService.ExecuteDatabaseOperation(DATABASE_NAME, DELETE_FILE_PATH, new DeleteEmbeddingByFileRequest(storeName, filePath), token);
+    public override Task DeleteEmbeddingByFile(string storeName, string filePath, string parentFileId, CancellationToken token) =>
+        rustService.ExecuteDatabaseOperation(DATABASE_NAME, DELETE_FILE_PATH, new DeleteEmbeddingByFileRequest(storeName, filePath, parentFileId), token);
+
+    public override Task RelocateEmbeddingFileAsync(string storeName, string parentFileId, string absolutePath, string relativePath, string fingerprint, CancellationToken token) =>
+        rustService.ExecuteDatabaseOperation(DATABASE_NAME, RELOCATE_FILE_PATH, new RelocateEmbeddingFileRequest(storeName, parentFileId, absolutePath, Path.GetFileName(absolutePath), relativePath, fingerprint), token);
 
     public override Task OptimizeVectorStore(string storeName, CancellationToken token) =>
         rustService.ExecuteDatabaseOperation(DATABASE_NAME, OPTIMIZE_PATH, new OptimizeVectorStoreRequest(storeName), token);
@@ -194,7 +198,9 @@ public sealed class QdrantEdgeClientImplementation(
     /// <param name="PointIds">The only points the search may return, or null for the whole store.</param>
     private sealed record SearchEmbeddingRequest(string StoreName, IReadOnlyList<float> Vector, int MaxMatches, IReadOnlyList<string>? PointIds);
 
-    private sealed record DeleteEmbeddingByFileRequest(string StoreName, string FilePath);
+    private sealed record DeleteEmbeddingByFileRequest(string StoreName, string FilePath, string ParentFileId);
+
+    private sealed record RelocateEmbeddingFileRequest(string StoreName, string ParentFileId, string AbsolutePath, string FileName, string RelativePath, string Fingerprint);
 
     private sealed record OptimizeVectorStoreRequest(string StoreName);
     

app/MindWork AI Studio/Tools/Databases/VectorStore/VectorStoreClient.cs

@@ -27,7 +27,12 @@ public abstract class VectorStoreClient(string name, string path): DatabaseClien
     /// <returns>The matches, best first, all of them inside the filter.</returns>
     public abstract Task<IReadOnlyList<VectorSearchResult>> SearchEmbeddingAsync(string storeName, IReadOnlyList<float> vector, int maxMatches, VectorSearchFilter filter, CancellationToken token);
 
-    public abstract Task DeleteEmbeddingByFile(string storeName, string filePath, CancellationToken token);
+    public abstract Task DeleteEmbeddingByFile(string storeName, string filePath, string parentFileId, CancellationToken token);
+
+    /// <summary>
+    /// Updates the location of a document's points without changing their ids or vectors.
+    /// </summary>
+    public abstract Task RelocateEmbeddingFileAsync(string storeName, string parentFileId, string absolutePath, string relativePath, string fingerprint, CancellationToken token);
 
     public abstract Task OptimizeVectorStore(string storeName, CancellationToken token);