From 7479db028a5e0b450394cba851bc61e8c4db4e3b Mon Sep 17 00:00:00 2001
From: ArzelaAscoIi <37148029+ArzelaAscoIi@users.noreply.github.com>
Date: Wed, 24 Apr 2024 14:49:38 +0200
Subject: [PATCH] fix: route not identified files to dead end edge (#7589)

* fix: route not identified files to dead end edge

* tests: fix unit tests

* tests: add file classifier pipeline run

* fix: typing
---
 haystack/nodes/file_classifier/file_type.py |  2 +-
 test/nodes/test_filetype_classifier.py      |  3 ++-
 test/pipelines/test_pipeline.py             | 21 +++++++++++++++++++++
 3 files changed, 24 insertions(+), 2 deletions(-)

diff --git a/haystack/nodes/file_classifier/file_type.py b/haystack/nodes/file_classifier/file_type.py
index 24b8caf460..67a3d8802b 100644
--- a/haystack/nodes/file_classifier/file_type.py
+++ b/haystack/nodes/file_classifier/file_type.py
@@ -134,7 +134,7 @@ def run(self, file_paths: Union[Path, List[Path], str, List[str], List[Union[Pat
                     paths[0],
                     self.supported_types,
                 )
-                return None, None
+                return {"file_paths": paths}, "output_dead_end"
             raise ValueError(
                 f"Files of type '{extension}' ({paths[0]}) are not supported. "
                 f"The supported types are: {self.supported_types}. "
diff --git a/test/nodes/test_filetype_classifier.py b/test/nodes/test_filetype_classifier.py
index e35c3c0cea..5b0f12c8da 100644
--- a/test/nodes/test_filetype_classifier.py
+++ b/test/nodes/test_filetype_classifier.py
@@ -176,7 +176,8 @@ def test_filetype_classifier_raise_on_error_disabled_unsupported_file_types(tmp_
     caplog.clear()
     with caplog.at_level(logging.WARNING):
         output, edge = node.run(test_file)
-        assert edge == output == None
+        assert edge == "output_dead_end"
+        assert output == {"file_paths": [test_file]}
         assert (
             f"Unsupported files of type '{file_type}' ({test_file!s}) found. Unsupported file types will be ignored"
             in caplog.text
diff --git a/test/pipelines/test_pipeline.py b/test/pipelines/test_pipeline.py
index 4713822b50..f9737f110e 100644
--- a/test/pipelines/test_pipeline.py
+++ b/test/pipelines/test_pipeline.py
@@ -1,3 +1,4 @@
+from pathlib import Path
 import ssl
 import json
 import platform
@@ -16,6 +17,7 @@
 from haystack.document_stores.deepsetcloud import DeepsetCloudDocumentStore
 from haystack.document_stores.elasticsearch import ElasticsearchDocumentStore
 from haystack.document_stores.memory import InMemoryDocumentStore
+from haystack.nodes.file_classifier.file_type import FileTypeClassifier
 from haystack.nodes.other.join_docs import JoinDocuments
 from haystack.nodes.base import BaseComponent
 from haystack.nodes.retriever.sparse import BM25Retriever
@@ -2191,6 +2193,25 @@ def test_pipeline_execution_using_join_preserves_changed_query():
     assert res["query"] == "This is a test."
 
 
+@pytest.mark.unit
+@pytest.mark.parametrize("file_type", ["csv", "xml"])
+def test_pipeline_execution_can_handle_unknown_edge_for_classifier(file_type: str) -> None:
+    """Tests running a classifier against an unexpected file type.
+
+    We need to route not expected file types against a dead end node.
+    Simply returning "None" for a classification does not work, since
+    the pipeline will raise an error, trying to route the result to the next node
+    here: https://github.com/deepset-ai/haystack/blob/b45ecb355636c3227185e97ee595006c06d17470/haystack/pipelines/base.py#L573
+
+    See fix pr: https://github.com/deepset-ai/haystack/pull/7589
+    """
+    classifier = FileTypeClassifier(raise_on_error=False)
+    pipeline = Pipeline()
+    pipeline.add_node(component=classifier, name="FileTypeClassifier", inputs=["File"])
+    res = pipeline.run_batch(file_paths=[f"./test.{file_type}"])
+    assert res["file_paths"] == [Path(f"./test.{file_type}")]
+
+
 @pytest.mark.unit
 def test_update_config_hash():
     fake_configs = {