feat: skip unknown file types [DC-1108] (#7542)

Co-authored-by: Marie-Luise Klaus <[email protected]>

haystack/nodes/file_classifier/file_type.py

@@ -26,7 +26,9 @@ class FileTypeClassifier(BaseComponent):
 
     outgoing_edges = len(DEFAULT_TYPES)
 
-    def __init__(self, supported_types: Optional[List[str]] = None, full_analysis: bool = False):
+    def __init__(
+        self, supported_types: Optional[List[str]] = None, full_analysis: bool = False, raise_on_error: bool = True
+    ):
         """
         Node that sends out files on a different output edge depending on their extension.
 
@@ -35,9 +37,11 @@ def __init__(self, supported_types: Optional[List[str]] = None, full_analysis: b
             You can't use lists with duplicate elements.
         :param full_analysis: If True, the whole file is analyzed to determine the file type.
             If False, only the first 2049 bytes are analyzed.
+        :param raise_on_error: If True, the node will raise an exception if the file type is not supported.
         """
         self.full_analysis = full_analysis
         self._default_types = False
+        self._raise_on_error = raise_on_error
         if supported_types is None:
             self._default_types = True
             supported_types = DEFAULT_TYPES
@@ -121,6 +125,16 @@ def run(self, file_paths: Union[Path, List[Path], str, List[str], List[Union[Pat
         try:
             index = self.supported_types.index(extension) + 1
         except ValueError:
+            if self._raise_on_error is False:
+                logger.warning(
+                    "Unsupported files of type '%s' (%s) found. "
+                    "Unsupported file types will be ignored during indexing as `raise_on_error` is set to `False`. "
+                    "The supported types are: %s. ",
+                    extension,
+                    paths[0],
+                    self.supported_types,
+                )
+                return None, None
             raise ValueError(
                 f"Files of type '{extension}' ({paths[0]}) are not supported. "
                 f"The supported types are: {self.supported_types}. "

test/nodes/test_filetype_classifier.py

@@ -166,3 +166,18 @@ def test_filetype_classifier_batched_same_media_extensions(tmp_path):
     output, edge = node.run_batch(test_files)
     assert edge == "output_1"
     assert output == {"file_paths": test_files}
+
+
+@pytest.mark.unit
+@pytest.mark.parametrize("file_type", ["csv", "json", "xml", "pptx", "xlsx"])
+def test_filetype_classifier_raise_on_error_disabled_unsupported_file_types(tmp_path, caplog, file_type):
+    node = FileTypeClassifier(raise_on_error=False)
+    test_file = tmp_path / f"test.{file_type}"
+    caplog.clear()
+    with caplog.at_level(logging.WARNING):
+        output, edge = node.run(test_file)
+        assert edge == output == None
+        assert (
+            f"Unsupported files of type '{file_type}' ({test_file!s}) found. Unsupported file types will be ignored"
+            in caplog.text
+        )