feat: implement CREATE INDEX statement and associated functionality

Author: srimon12

src/qql/ast_nodes.py

@@ -145,6 +145,13 @@ class CreateCollectionStmt:
     model: str | None = None  # dense model; None → use config default
 
 
+@dataclass(frozen=True)
+class CreateIndexStmt:
+    collection: str
+    field_name: str
+    schema: str
+
+
 @dataclass(frozen=True)
 class DropCollectionStmt:
     collection: str
@@ -188,14 +195,16 @@ class RecommendStmt:
 @dataclass(frozen=True)
 class DeleteStmt:
     collection: str
-    point_id: str | int
+    point_id: str | int | None = None
+    query_filter: FilterExpr | None = None
 
 
 # Union type for all top-level statement nodes
 ASTNode = (
     InsertStmt
     | InsertBulkStmt
     | CreateCollectionStmt
+    | CreateIndexStmt
     | DropCollectionStmt
     | ShowCollectionsStmt
     | SearchStmt

src/qql/executor.py

@@ -26,6 +26,7 @@
     MatchValue,
     Modifier,
     PayloadField,
+    PayloadSchemaType,
     PointStruct,
     Prefetch,
     Range,
@@ -44,6 +45,7 @@
     BetweenExpr,
     CompareExpr,
     CreateCollectionStmt,
+    CreateIndexStmt,
     DeleteStmt,
     DropCollectionStmt,
     FilterExpr,
@@ -93,6 +95,8 @@ def execute(self, node: ASTNode) -> ExecutionResult:
             return self._execute_insert(node)
         if isinstance(node, CreateCollectionStmt):
             return self._execute_create(node)
+        if isinstance(node, CreateIndexStmt):
+            return self._execute_create_index(node)
         if isinstance(node, DropCollectionStmt):
             return self._execute_drop(node)
         if isinstance(node, ShowCollectionsStmt):
@@ -321,6 +325,43 @@ def _execute_create(self, node: CreateCollectionStmt) -> ExecutionResult:
             message=f"Collection '{node.collection}' created ({dims}-dimensional vectors, cosine distance)",
         )
 
+    def _execute_create_index(self, node: CreateIndexStmt) -> ExecutionResult:
+        if not self._client.collection_exists(node.collection):
+            raise QQLRuntimeError(f"Collection '{node.collection}' does not exist")
+
+        schema_map = {
+            "keyword": PayloadSchemaType.KEYWORD,
+            "integer": PayloadSchemaType.INTEGER,
+            "float": PayloadSchemaType.FLOAT,
+            "bool": PayloadSchemaType.BOOL,
+            "text": PayloadSchemaType.TEXT,
+            "geo": PayloadSchemaType.GEO,
+            "datetime": PayloadSchemaType.DATETIME,
+        }
+        try:
+            field_schema = schema_map[node.schema]
+        except KeyError as e:
+            raise QQLRuntimeError(
+                "Unknown index type '"
+                f"{node.schema}'. Expected one of: keyword, integer, float, bool, text, geo, datetime"
+            ) from e
+
+        try:
+            self._client.create_payload_index(
+                collection_name=node.collection,
+                field_name=node.field_name,
+                field_schema=field_schema,
+            )
+        except UnexpectedResponse as e:
+            raise QQLRuntimeError(f"Qdrant error during CREATE INDEX: {e}") from e
+
+        return ExecutionResult(
+            success=True,
+            message=(
+                f"Created index on '{node.collection}.{node.field_name}' as '{node.schema}'"
+            ),
+        )
+
     def _execute_drop(self, node: DropCollectionStmt) -> ExecutionResult:
         if not self._client.collection_exists(node.collection):
             raise QQLRuntimeError(f"Collection '{node.collection}' does not exist")
@@ -648,9 +689,25 @@ def _execute_delete(self, node: DeleteStmt) -> ExecutionResult:
         if not self._client.collection_exists(node.collection):
             raise QQLRuntimeError(f"Collection '{node.collection}' does not exist")
 
-        from qdrant_client.models import PointIdsList
-
         try:
+            if node.query_filter is not None:
+                self._client.delete(
+                    collection_name=node.collection,
+                    wait=True,
+                    points_selector=self._wrap_as_filter(
+                        self._build_qdrant_filter(node.query_filter)
+                    ),
+                )
+                return ExecutionResult(
+                    success=True,
+                    message=f"Deleted points from '{node.collection}' by filter",
+                )
+
+            from qdrant_client.models import PointIdsList
+
+            if node.point_id is None:
+                raise QQLRuntimeError("DELETE requires either a point id or a filter")
+
             self._client.delete(
                 collection_name=node.collection,
                 wait=True,

src/qql/lexer.py

@@ -21,6 +21,8 @@ class TokenKind(Enum):
     WITH = auto()
     ACORN = auto()
     CREATE = auto()
+    INDEX = auto()
+    ON = auto()
     DROP = auto()
     SHOW = auto()
     COLLECTIONS = auto()
@@ -42,6 +44,8 @@ class TokenKind(Enum):
     FROM = auto()
     WHERE = auto()
     ID = auto()
+    FOR = auto()
+    TYPE = auto()
     # ── Filter keywords ───────────────────────────────────────────────────
     AND = auto()
     OR = auto()
@@ -96,6 +100,8 @@ class TokenKind(Enum):
     "WITH": TokenKind.WITH,
     "ACORN": TokenKind.ACORN,
     "CREATE": TokenKind.CREATE,
+    "INDEX": TokenKind.INDEX,
+    "ON": TokenKind.ON,
     "DROP": TokenKind.DROP,
     "SHOW": TokenKind.SHOW,
     "COLLECTIONS": TokenKind.COLLECTIONS,
@@ -117,6 +123,8 @@ class TokenKind(Enum):
     "FROM": TokenKind.FROM,
     "WHERE": TokenKind.WHERE,
     "ID": TokenKind.ID,
+    "FOR": TokenKind.FOR,
+    "TYPE": TokenKind.TYPE,
     # Filter keywords
     "AND": TokenKind.AND,
     "OR": TokenKind.OR,

src/qql/parser.py

@@ -6,6 +6,7 @@
     BetweenExpr,
     CompareExpr,
     CreateCollectionStmt,
+    CreateIndexStmt,
     DeleteStmt,
     DropCollectionStmt,
     FilterExpr,
@@ -150,34 +151,45 @@ def _parse_insert_bulk_body(self) -> InsertBulkStmt:
 
     def _parse_create(self) -> CreateCollectionStmt:
         self._expect(TokenKind.CREATE)
-        self._expect(TokenKind.COLLECTION)
-        collection = self._parse_identifier()
-        hybrid: bool = False
-        model: str | None = None
-
-        if self._peek().kind == TokenKind.HYBRID:
-            # Bare HYBRID shorthand — backward compat
+        if self._peek().kind == TokenKind.COLLECTION:
             self._advance()
-            hybrid = True
-        elif self._peek().kind == TokenKind.USING:
-            self._advance()  # consume USING
+            collection = self._parse_identifier()
+            hybrid: bool = False
+            model: str | None = None
+
             if self._peek().kind == TokenKind.HYBRID:
-                self._advance()  # consume HYBRID
+                # Bare HYBRID shorthand — backward compat
+                self._advance()
                 hybrid = True
-                # Optional DENSE MODEL sub-clause
-                if self._peek().kind == TokenKind.DENSE:
-                    self._advance()  # consume DENSE
+            elif self._peek().kind == TokenKind.USING:
+                self._advance()  # consume USING
+                if self._peek().kind == TokenKind.HYBRID:
+                    self._advance()  # consume HYBRID
+                    hybrid = True
+                    # Optional DENSE MODEL sub-clause
+                    if self._peek().kind == TokenKind.DENSE:
+                        self._advance()  # consume DENSE
+                        self._expect(TokenKind.MODEL)
+                        model = self._expect(TokenKind.STRING).value
+                else:
                     self._expect(TokenKind.MODEL)
                     model = self._expect(TokenKind.STRING).value
-            else:
-                self._expect(TokenKind.MODEL)
-                model = self._expect(TokenKind.STRING).value
 
-        return CreateCollectionStmt(
-            collection=collection,
-            hybrid=hybrid,
-            model=model,
-        )
+            return CreateCollectionStmt(
+                collection=collection,
+                hybrid=hybrid,
+                model=model,
+            )
+
+        self._expect(TokenKind.INDEX)
+        self._expect(TokenKind.ON)
+        self._expect(TokenKind.COLLECTION)
+        collection = self._parse_identifier()
+        self._expect(TokenKind.FOR)
+        field_name = self._parse_field_path()
+        self._expect(TokenKind.TYPE)
+        schema = self._expect(TokenKind.IDENTIFIER).value.lower()
+        return CreateIndexStmt(collection=collection, field_name=field_name, schema=schema)
 
     def _parse_drop(self) -> DropCollectionStmt:
         self._expect(TokenKind.DROP)
@@ -356,20 +368,24 @@ def _parse_delete(self) -> DeleteStmt:
         self._expect(TokenKind.FROM)
         collection = self._parse_identifier()
         self._expect(TokenKind.WHERE)
-        self._expect(TokenKind.ID)
-        self._expect(TokenKind.EQUALS)
-        tok = self._peek()
-        if tok.kind == TokenKind.STRING:
-            self._advance()
-            point_id: str | int = tok.value
-        elif tok.kind == TokenKind.INTEGER:
+        if self._peek().kind == TokenKind.ID:
             self._advance()
-            point_id = int(tok.value)
-        else:
-            raise QQLSyntaxError(
-                f"Expected string or integer for point id, got '{tok.value}'", tok.pos
-            )
-        return DeleteStmt(collection=collection, point_id=point_id)
+            self._expect(TokenKind.EQUALS)
+            tok = self._peek()
+            if tok.kind == TokenKind.STRING:
+                self._advance()
+                point_id: str | int = tok.value
+            elif tok.kind == TokenKind.INTEGER:
+                self._advance()
+                point_id = int(tok.value)
+            else:
+                raise QQLSyntaxError(
+                    f"Expected string or integer for point id, got '{tok.value}'", tok.pos
+                )
+            return DeleteStmt(collection=collection, point_id=point_id)
+
+        query_filter = self._parse_filter_expr()
+        return DeleteStmt(collection=collection, query_filter=query_filter)
 
     # ── WHERE clause filter parsing (precedence: NOT > AND > OR) ─────────
 

tests/test_executor.py

@@ -2,6 +2,7 @@
 
 from qql.ast_nodes import (
     CreateCollectionStmt,
+    CreateIndexStmt,
     DeleteStmt,
     DropCollectionStmt,
     InsertBulkStmt,
@@ -246,6 +247,21 @@ def test_create_existing_collection_is_noop(self, executor, mock_client):
         assert "already exists" in result.message
 
 
+class TestCreateIndex:
+    def test_create_index_calls_qdrant(self, executor, mock_client):
+        mock_client.collection_exists.return_value = True
+        node = CreateIndexStmt(collection="articles", field_name="category", schema="keyword")
+        result = executor.execute(node)
+        mock_client.create_payload_index.assert_called_once()
+        assert result.success is True
+
+    def test_create_index_nonexistent_collection_raises(self, executor, mock_client):
+        mock_client.collection_exists.return_value = False
+        node = CreateIndexStmt(collection="ghost", field_name="category", schema="keyword")
+        with pytest.raises(QQLRuntimeError, match="does not exist"):
+            executor.execute(node)
+
+
 class TestCreateWithModel:
     def test_create_with_model_passes_model_to_embedder(self, mock_client, cfg, mocker):
         mock_emb = mocker.MagicMock()
@@ -630,6 +646,20 @@ def test_delete_calls_qdrant_delete(self, executor, mock_client):
         mock_client.delete.assert_called_once()
         assert result.success is True
 
+    def test_delete_by_filter_calls_qdrant_delete_with_filter(self, executor, mock_client):
+        from qdrant_client.models import Filter
+        from qql.ast_nodes import CompareExpr
+
+        mock_client.collection_exists.return_value = True
+        node = DeleteStmt(
+            collection="articles",
+            query_filter=CompareExpr(field="category", op="=", value="archived"),
+        )
+        result = executor.execute(node)
+        selector = mock_client.delete.call_args.kwargs["points_selector"]
+        assert isinstance(selector, Filter)
+        assert result.success is True
+
     def test_delete_nonexistent_collection_raises(self, executor, mock_client):
         mock_client.collection_exists.return_value = False
         node = DeleteStmt(collection="ghost", point_id="x")

tests/test_parser.py

@@ -5,6 +5,7 @@
     BetweenExpr,
     CompareExpr,
     CreateCollectionStmt,
+    CreateIndexStmt,
     DeleteStmt,
     DropCollectionStmt,
     InExpr,
@@ -165,6 +166,13 @@ def test_create_collection(self):
         assert isinstance(node, CreateCollectionStmt)
         assert node.collection == "my_col"
 
+    def test_create_index(self):
+        node = parse("CREATE INDEX ON COLLECTION articles FOR category TYPE keyword")
+        assert isinstance(node, CreateIndexStmt)
+        assert node.collection == "articles"
+        assert node.field_name == "category"
+        assert node.schema == "keyword"
+
 
 class TestDrop:
     def test_drop_collection(self):
@@ -199,12 +207,21 @@ def test_delete_by_string_id(self):
         assert isinstance(node, DeleteStmt)
         assert node.collection == "notes"
         assert node.point_id == "abc-123"
+        assert node.query_filter is None
 
     def test_delete_by_integer_id(self):
         node = parse("DELETE FROM notes WHERE id = 99")
         assert isinstance(node, DeleteStmt)
         assert node.point_id == 99
 
+    def test_delete_by_filter(self):
+        node = parse("DELETE FROM articles WHERE category = 'archived'")
+        assert isinstance(node, DeleteStmt)
+        assert node.point_id is None
+        assert isinstance(node.query_filter, CompareExpr)
+        assert node.query_filter.field == "category"
+        assert node.query_filter.value == "archived"
+
 
 class TestRecommend:
     def test_recommend_with_positive_ids(self):