fix(tools): isolate RAG auto-detection across items and support uppercase extensions

lib/crewai-tools/src/crewai_tools/adapters/crewai_rag_adapter.py

@@ -197,7 +197,8 @@ def add(self, *args: ContentItem, **kwargs: Unpack[AddDocumentParams]) -> None:
             else:
                 source_ref = str(arg)
 
-            if not data_type:
+            item_data_type = data_type
+            if not item_data_type:
                 ext = os.path.splitext(source_ref)[1].lower()
                 is_url = source_ref.startswith(("http://", "https://", "file://"))
                 if (
@@ -206,9 +207,9 @@ def add(self, *args: ContentItem, **kwargs: Unpack[AddDocumentParams]) -> None:
                     and not os.path.isfile(source_ref)
                 ):
                     raise FileNotFoundError(f"File does not exist: {source_ref}")
-                data_type = DataTypes.from_content(source_ref)
+                item_data_type = DataTypes.from_content(source_ref)
 
-            if data_type == DataType.DIRECTORY:
+            if item_data_type == DataType.DIRECTORY:
                 if not os.path.isdir(source_ref):
                     raise ValueError(f"Directory does not exist: {source_ref}")
 
@@ -304,7 +305,7 @@ def add(self, *args: ContentItem, **kwargs: Unpack[AddDocumentParams]) -> None:
                 metadata: dict[str, Any] = base_metadata.copy()
                 source_content = SourceContent(source_ref)
 
-                if data_type in [
+                if item_data_type in [
                     DataType.PDF_FILE,
                     DataType.TEXT_FILE,
                     DataType.DOCX,
@@ -316,8 +317,8 @@ def add(self, *args: ContentItem, **kwargs: Unpack[AddDocumentParams]) -> None:
                     if not source_content.is_url() and not source_content.path_exists():
                         raise FileNotFoundError(f"File does not exist: {source_ref}")
 
-                loader = data_type.get_loader()
-                chunker = data_type.get_chunker()
+                loader = item_data_type.get_loader()
+                chunker = item_data_type.get_chunker()
 
                 loader_result: LoaderResult = loader.load(source_content, **kwargs)
 
@@ -326,7 +327,7 @@ def add(self, *args: ContentItem, **kwargs: Unpack[AddDocumentParams]) -> None:
                 for i, chunk in enumerate(chunks):
                     chunk_metadata: dict[str, Any] = metadata.copy()
                     chunk_metadata.update(loader_result.metadata)
-                    chunk_metadata["data_type"] = str(data_type)
+                    chunk_metadata["data_type"] = str(item_data_type)
                     chunk_metadata["chunk_index"] = i
                     chunk_metadata["total_chunks"] = len(chunks)
                     chunk_metadata["source"] = source_ref

lib/crewai-tools/src/crewai_tools/rag/data_types.py

@@ -123,8 +123,9 @@ def get_file_type(path: str) -> DataType | None:
                 ".xml": DataType.XML,
                 ".txt": DataType.TEXT_FILE,
             }
+            path_lower = path.lower()
             for ext, dtype in mapping.items():
-                if path.endswith(ext):
+                if path_lower.endswith(ext):
                     return dtype
             return None
 

lib/crewai-tools/tests/tools/rag/test_rag_tool_add_data_type.py

@@ -475,4 +475,54 @@ def test_metadata_passed_to_documents(
             assert any(
                 doc.get("metadata", {}).get("custom_key") == "custom_value"
                 for doc in documents
-            )
+            )
+
+
+class TestAutoDetectMultipleItemsAndCaseSensitivity:
+    """Tests for case-insensitive extensions and multi-item auto-detection."""
+
+    def test_auto_detect_uppercase_file_extension(
+        self, rag_tool: RagTool, mock_rag_client: MagicMock
+    ) -> None:
+        """Test that uppercase file extensions like .JSON and .CSV are correctly detected."""
+        with TemporaryDirectory() as tmpdir:
+            json_file = Path(tmpdir) / "DATA.JSON"
+            json_file.write_text('{"name": "test"}')
+
+            rag_tool.add(str(json_file))
+
+            assert mock_rag_client.add_documents.called
+            call_args = mock_rag_client.add_documents.call_args
+            documents = call_args.kwargs.get("documents", [])
+            assert any(
+                doc.get("metadata", {}).get("data_type") == str(DataType.JSON)
+                for doc in documents
+            )
+
+    def test_auto_detect_uppercase_url_extension(self) -> None:
+        """Test that URLs with uppercase extensions resolve to their correct DataType."""
+        from crewai_tools.rag.data_types import DataTypes
+
+        assert DataTypes.from_content("https://example.com/report.PDF") == DataType.PDF_FILE
+        assert DataTypes.from_content("https://example.com/data.CSV") == DataType.CSV
+
+    def test_auto_detect_multiple_items_independent_types(
+        self, rag_tool: RagTool, mock_rag_client: MagicMock
+    ) -> None:
+        """Test that adding a directory followed by a file does not leak DataType.DIRECTORY."""
+        with TemporaryDirectory() as tmpdir:
+            subfolder = Path(tmpdir) / "subfolder"
+            subfolder.mkdir()
+            (subfolder / "inner.txt").write_text("inner content")
+
+            file2 = Path(tmpdir) / "outer.txt"
+            file2.write_text("outer content")
+
+            # Must not raise ValueError: Directory does not exist on outer.txt
+            rag_tool.add(str(subfolder), str(file2))
+
+            assert mock_rag_client.add_documents.called
+            call_args = mock_rag_client.add_documents.call_args
+            documents = call_args.kwargs.get("documents", [])
+            sources = [doc.get("metadata", {}).get("source") for doc in documents]
+            assert any("outer.txt" in s for s in sources if s)