Add category field to dataset metadata for AI-generated dataset isolation

Adds a `category` column to the `_datasets_meta` schema via a new migration,
threaded through the full Rust core → FFI → SDK → Desktop stack. The
ChartDatasetGenerator now tags AI-derived datasets with category "ai-generated"
when calling ImportFromContentAsync. The SDK's IDatasetService and DatasetInfo
types expose the new optional category field with backward-compatible defaults.

Rust changes: schema migration, DatasetMeta.category field, store CRUD/mutations
updated to read/write/copy category, new set_category() helper. FFI request
structs (CreateEmptyRequest, ImportContentRequest) accept optional category.

SDK/Desktop: DatasetInfo gains Category property, IDatasetService methods
CreateEmptyDatasetAsync and ImportFromContentAsync accept optional category
parameter, DatasetService.Mutations threads it to FFI JSON requests.

Version bumps: SDK 1.55.0 → 1.56.0, Desktop 1.55.5 → 1.56.0.

Author: skmelendez

core/privstack-datasets/src/schema.rs

@@ -74,6 +74,10 @@ CREATE TABLE IF NOT EXISTS _dataset_saved_queries (
 const SAVED_QUERIES_ADD_IS_VIEW: &str =
     "ALTER TABLE _dataset_saved_queries ADD COLUMN IF NOT EXISTS is_view BOOLEAN DEFAULT FALSE;";
 
+/// Migration: add category column to datasets metadata for AI-generated dataset isolation.
+const DATASETS_META_ADD_CATEGORY: &str =
+    "ALTER TABLE _datasets_meta ADD COLUMN IF NOT EXISTS category VARCHAR;";
+
 /// Initialize all dataset schema tables.
 pub fn initialize_datasets_schema(conn: &Connection) -> DatasetResult<()> {
     conn.execute_batch(DATASETS_META_DDL)?;
@@ -83,6 +87,7 @@ pub fn initialize_datasets_schema(conn: &Connection) -> DatasetResult<()> {
     conn.execute_batch(DATASET_SAVED_QUERIES_DDL)?;
     // Migrations
     conn.execute_batch(SAVED_QUERIES_ADD_IS_VIEW)?;
+    conn.execute_batch(DATASETS_META_ADD_CATEGORY)?;
     Ok(())
 }
 

core/privstack-datasets/src/store/crud.rs

@@ -67,6 +67,7 @@ impl DatasetStore {
             source_file_name,
             row_count,
             columns,
+            category: None,
             created_at: now,
             modified_at: now,
         })
@@ -76,7 +77,7 @@ impl DatasetStore {
     pub fn list(&self) -> DatasetResult<Vec<DatasetMeta>> {
         let conn = self.lock_conn();
         let mut stmt = conn.prepare(
-            "SELECT id, name, source_file_name, row_count, columns_json, created_at, modified_at FROM _datasets_meta ORDER BY modified_at DESC"
+            "SELECT id, name, source_file_name, row_count, columns_json, created_at, modified_at, category FROM _datasets_meta ORDER BY modified_at DESC"
         )?;
 
         let rows = stmt
@@ -89,6 +90,7 @@ impl DatasetStore {
                     row.get::<_, String>(4)?,
                     row.get::<_, i64>(5)?,
                     row.get::<_, i64>(6)?,
+                    row.get::<_, Option<String>>(7)?,
                 ))
             })?
             .filter_map(|r| r.ok())
@@ -98,7 +100,7 @@ impl DatasetStore {
         drop(conn);
 
         rows.into_iter()
-            .map(|(id, name, source, row_count, cols_json, created, modified)| {
+            .map(|(id, name, source, row_count, cols_json, created, modified, category)| {
                 let columns: Vec<DatasetColumn> =
                     serde_json::from_str(&cols_json).unwrap_or_default();
                 Ok(DatasetMeta {
@@ -109,6 +111,7 @@ impl DatasetStore {
                     source_file_name: source,
                     row_count,
                     columns,
+                    category,
                     created_at: created,
                     modified_at: modified,
                 })
@@ -120,7 +123,7 @@ impl DatasetStore {
     pub fn get(&self, id: &DatasetId) -> DatasetResult<DatasetMeta> {
         let conn = self.lock_conn();
         let result = conn.query_row(
-            "SELECT name, source_file_name, row_count, columns_json, created_at, modified_at FROM _datasets_meta WHERE id = ?",
+            "SELECT name, source_file_name, row_count, columns_json, created_at, modified_at, category FROM _datasets_meta WHERE id = ?",
             params![id.to_string()],
             |row| {
                 Ok((
@@ -130,12 +133,13 @@ impl DatasetStore {
                     row.get::<_, String>(3)?,
                     row.get::<_, i64>(4)?,
                     row.get::<_, i64>(5)?,
+                    row.get::<_, Option<String>>(6)?,
                 ))
             },
         );
 
         match result {
-            Ok((name, source, row_count, cols_json, created, modified)) => {
+            Ok((name, source, row_count, cols_json, created, modified, category)) => {
                 let columns: Vec<DatasetColumn> =
                     serde_json::from_str(&cols_json).unwrap_or_default();
                 Ok(DatasetMeta {
@@ -144,6 +148,7 @@ impl DatasetStore {
                     source_file_name: source,
                     row_count,
                     columns,
+                    category,
                     created_at: created,
                     modified_at: modified,
                 })
@@ -188,6 +193,21 @@ impl DatasetStore {
         Ok(())
     }
 
+    /// Set or clear the category for a dataset.
+    pub fn set_category(&self, id: &DatasetId, category: Option<&str>) -> DatasetResult<()> {
+        let now = now_millis();
+        let conn = self.lock_conn();
+        let updated = conn.execute(
+            "UPDATE _datasets_meta SET category = ?, modified_at = ? WHERE id = ?",
+            params![category, now, id.to_string()],
+        )?;
+
+        if updated == 0 {
+            return Err(DatasetError::NotFound(id.to_string()));
+        }
+        Ok(())
+    }
+
     /// Rename a dataset.
     pub fn rename(&self, id: &DatasetId, new_name: &str) -> DatasetResult<()> {
         let now = now_millis();

core/privstack-datasets/src/store/mutations.rs

@@ -14,6 +14,7 @@ impl DatasetStore {
         &self,
         name: &str,
         columns: &[ColumnDef],
+        category: Option<&str>,
     ) -> DatasetResult<DatasetMeta> {
         if columns.is_empty() {
             return Err(DatasetError::InvalidQuery(
@@ -44,9 +45,9 @@ impl DatasetStore {
         let columns_json = serde_json::to_string(&ds_columns)?;
 
         conn.execute(
-            r#"INSERT INTO _datasets_meta (id, name, source_file_name, row_count, columns_json, created_at, modified_at)
-               VALUES (?, ?, NULL, 0, ?, ?, ?)"#,
-            params![id.to_string(), name, columns_json, now, now],
+            r#"INSERT INTO _datasets_meta (id, name, source_file_name, row_count, columns_json, category, created_at, modified_at)
+               VALUES (?, ?, NULL, 0, ?, ?, ?, ?)"#,
+            params![id.to_string(), name, columns_json, category, now, now],
         )?;
 
         info!(dataset_id = %id, name, "Empty dataset created");
@@ -57,6 +58,7 @@ impl DatasetStore {
             source_file_name: None,
             row_count: 0,
             columns: ds_columns,
+            category: category.map(|s| s.to_string()),
             created_at: now,
             modified_at: now,
         })
@@ -75,6 +77,15 @@ impl DatasetStore {
 
         let conn = self.lock_conn();
 
+        // Read source category before duplicating
+        let source_category: Option<String> = conn
+            .query_row(
+                "SELECT category FROM _datasets_meta WHERE id = ?",
+                params![source_id.to_string()],
+                |row| row.get(0),
+            )
+            .unwrap_or(None);
+
         let create_sql = format!("CREATE TABLE {new_table} AS SELECT * FROM {source_table}");
         conn.execute_batch(&create_sql).map_err(|e| {
             DatasetError::ImportFailed(format!("Failed to duplicate dataset: {e}"))
@@ -88,9 +99,9 @@ impl DatasetStore {
         let columns_json = serde_json::to_string(&columns)?;
 
         conn.execute(
-            r#"INSERT INTO _datasets_meta (id, name, source_file_name, row_count, columns_json, created_at, modified_at)
-               VALUES (?, ?, NULL, ?, ?, ?, ?)"#,
-            params![new_id.to_string(), new_name, row_count, columns_json, now, now],
+            r#"INSERT INTO _datasets_meta (id, name, source_file_name, row_count, columns_json, category, created_at, modified_at)
+               VALUES (?, ?, NULL, ?, ?, ?, ?, ?)"#,
+            params![new_id.to_string(), new_name, row_count, columns_json, source_category, now, now],
         )?;
 
         info!(dataset_id = %new_id, new_name, "Dataset duplicated from {}", source_id);
@@ -101,6 +112,7 @@ impl DatasetStore {
             source_file_name: None,
             row_count,
             columns,
+            category: source_category,
             created_at: now,
             modified_at: now,
         })
@@ -111,6 +123,7 @@ impl DatasetStore {
         &self,
         csv_content: &str,
         name: &str,
+        category: Option<&str>,
     ) -> DatasetResult<DatasetMeta> {
         let id = DatasetId::new();
         let table = dataset_table_name(&id);
@@ -143,9 +156,9 @@ impl DatasetStore {
         let columns_json = serde_json::to_string(&columns)?;
 
         conn.execute(
-            r#"INSERT INTO _datasets_meta (id, name, source_file_name, row_count, columns_json, created_at, modified_at)
-               VALUES (?, ?, NULL, ?, ?, ?, ?)"#,
-            params![id.to_string(), name, row_count, columns_json, now, now],
+            r#"INSERT INTO _datasets_meta (id, name, source_file_name, row_count, columns_json, category, created_at, modified_at)
+               VALUES (?, ?, NULL, ?, ?, ?, ?, ?)"#,
+            params![id.to_string(), name, row_count, columns_json, category, now, now],
         )?;
 
         info!(dataset_id = %id, name, row_count, "Dataset imported from content");
@@ -156,6 +169,7 @@ impl DatasetStore {
             source_file_name: None,
             row_count,
             columns,
+            category: category.map(|s| s.to_string()),
             created_at: now,
             modified_at: now,
         })

core/privstack-datasets/src/types.rs

@@ -80,6 +80,7 @@ pub struct DatasetMeta {
     pub source_file_name: Option<String>,
     pub row_count: i64,
     pub columns: Vec<DatasetColumn>,
+    pub category: Option<String>,
     pub created_at: i64,
     pub modified_at: i64,
 }

core/privstack-ffi/src/datasets/mod.rs

@@ -175,6 +175,7 @@ pub(crate) struct SavedQueryRequest {
 pub(crate) struct CreateEmptyRequest {
     pub name: String,
     pub columns: Vec<privstack_datasets::ColumnDef>,
+    pub category: Option<String>,
 }
 
 #[derive(Deserialize)]
@@ -187,6 +188,7 @@ pub(crate) struct DuplicateRequest {
 pub(crate) struct ImportContentRequest {
     pub content: String,
     pub name: String,
+    pub category: Option<String>,
 }
 
 #[derive(Deserialize)]

core/privstack-ffi/src/datasets/mutations.rs

@@ -18,7 +18,7 @@ pub unsafe extern "C" fn privstack_dataset_create_empty(
         let req = parse_json_request!(request_json, CreateEmptyRequest);
 
         with_store_json!(r#"{"error":"not initialized"}"#, |store| {
-            match store.create_empty(&req.name, &req.columns) {
+            match store.create_empty(&req.name, &req.columns, req.category.as_deref()) {
                 Ok(meta) => {
                     let json =
                         serde_json::to_string(&meta).unwrap_or_else(|_| "{}".to_string());
@@ -73,7 +73,7 @@ pub unsafe extern "C" fn privstack_dataset_import_content(
         let req = parse_json_request!(request_json, ImportContentRequest);
 
         with_store_json!(r#"{"error":"not initialized"}"#, |store| {
-            match store.import_csv_content(&req.content, &req.name) {
+            match store.import_csv_content(&req.content, &req.name, req.category.as_deref()) {
                 Ok(meta) => {
                     let json =
                         serde_json::to_string(&meta).unwrap_or_else(|_| "{}".to_string());

desktop/Directory.Build.props

@@ -1,5 +1,5 @@
 <Project>
   <PropertyGroup>
-    <PrivStackSdkVersion>1.55.0</PrivStackSdkVersion>
+    <PrivStackSdkVersion>1.56.0</PrivStackSdkVersion>
   </PropertyGroup>
 </Project>

desktop/PrivStack.Desktop/PrivStack.Desktop.csproj

@@ -15,7 +15,7 @@
     <CFBundleDisplayName>PrivStack</CFBundleDisplayName>
     <AssemblyTitle>PrivStack</AssemblyTitle>
     <Product>PrivStack</Product>
-    <Version>1.55.5</Version>
+    <Version>1.56.0</Version>
 
     <!-- LlamaSharp.Backend.Cpu ships multiple AVX variants (avx/avx2/avx512/base) that collide on publish -->
     <ErrorOnDuplicatePublishOutputFiles>false</ErrorOnDuplicatePublishOutputFiles>

desktop/PrivStack.Desktop/Services/AI/ChartDatasetGenerator.cs

@@ -54,7 +54,7 @@ public ChartDatasetGenerator(IDatasetService datasetService)
                 return null;
 
             var datasetName = $"[AI] {SanitizeName(chart.Title)} — {sourceDatasetName}";
-            var dataset = await _datasetService.ImportFromContentAsync(csv, datasetName, ct);
+            var dataset = await _datasetService.ImportFromContentAsync(csv, datasetName, ct, category: "ai-generated");
             Log.Information("Created chart dataset {Id} ({Name}) with aggregated data", dataset.Id, datasetName);
             return dataset.Id.Value;
         }

desktop/PrivStack.Desktop/Services/DatasetService.Mutations.cs

@@ -11,9 +11,9 @@ public sealed partial class DatasetService
     // ── Dataset Creation ────────────────────────────────────────────────
 
     public Task<DatasetInfo> CreateEmptyDatasetAsync(
-        string name, IReadOnlyList<DatasetColumnDef> columns, CancellationToken ct)
+        string name, IReadOnlyList<DatasetColumnDef> columns, CancellationToken ct, string? category = null)
     {
-        var request = new { name, columns };
+        var request = new { name, columns, category };
         var ptr = DatasetNative.CreateEmpty(JsonSerializer.Serialize(request, JsonOptions));
         try
         {
@@ -58,9 +58,9 @@ public Task<DatasetInfo> DuplicateDatasetAsync(
         }
     }
 
-    public Task<DatasetInfo> ImportFromContentAsync(string content, string name, CancellationToken ct)
+    public Task<DatasetInfo> ImportFromContentAsync(string content, string name, CancellationToken ct, string? category = null)
     {
-        var request = new { content, name };
+        var request = new { content, name, category };
         var ptr = DatasetNative.ImportContent(JsonSerializer.Serialize(request, JsonOptions));
         try
         {