fix(asr): Implement robust status check in AsrPipeline (#2442)

* test: Add failing test case for silent audio file

* fix: Implement robust status check in AsrPipeline

* DCO Remediation Commit for mastermaxx03 <srivastavaanimesh22@gmail.com>I, mastermaxx03 <srivastavaanimesh22@gmail.com>, hereby add my Signed-off-by to this commit: 5fc4d512b330bb0cd347da4cbcca0fbe9687898aI, mastermaxx03 <srivastavaanimesh22@gmail.com>, hereby add my Signed-off-by to this commit: 31a4e9a5f1

Signed-off-by: mastermaxx03 <srivastavaanimesh22@gmail.com>

* DCO Remediation Commit for mastermaxx03 <srivastavaanimesh22@gmail.com>

I, mastermaxx03 <srivastavaanimesh22@gmail.com>, hereby add my Signed-off-by to this commit: 5fc4d512b3
I, mastermaxx03 <srivastavaanimesh22@gmail.com>, hereby add my Signed-off-by to this commit: 31a4e9a5f1

Signed-off-by: mastermaxx03 <srivastavaanimesh22@gmail.com>

* DCO Remediation Commit for mastermaxx03 <srivastavaanimesh22@gmail.com>

I, mastermaxx03 <srivastavaanimesh22@gmail.com>, hereby add my Signed-off-by to this commit: 5fc4d512b3
I, mastermaxx03 <srivastavaanimesh22@gmail.com>, hereby add my Signed-off-by to this commit: 31a4e9a5f1

Signed-off-by: mastermaxx03 <srivastavaanimesh22@gmail.com>

---------

Signed-off-by: mastermaxx03 <srivastavaanimesh22@gmail.com>

docling/pipeline/asr_pipeline.py

@@ -186,7 +186,6 @@ class _NativeWhisperModel:
                     label=DocItemLabel.TEXT, text=citem.to_string()
                 )
 
-            conv_res.status = ConversionStatus.SUCCESS
             return conv_res
 
         except Exception as exc:
@@ -249,9 +248,29 @@ class AsrPipeline(BasePipeline):
         else:
             _log.error(f"No model support for {self.pipeline_options.asr_options}")
 
+    def _has_text(self, document: "DoclingDocument") -> bool:
+        """
+        Helper method to check if the document contains any transcribed text.
+        A transcription is considered non-empty if the .texts list contains items with actual, non whitespace content.
+        """
+        if not document or not document.texts:
+            return False
+        for item in document.texts:
+            if item.text and item.text.strip():
+                return True
+        return False
+
     def _determine_status(self, conv_res: ConversionResult) -> ConversionStatus:
-        status = ConversionStatus.SUCCESS
-        return status
+        """Determines the final status of ASR Conversion based on its result."""
+        if conv_res.status == ConversionStatus.FAILURE or conv_res.errors:
+            return ConversionStatus.FAILURE
+        if not self._has_text(conv_res.document):
+            _log.warning(
+                "ASR conversion resulted in an empty document."
+                f"File: {conv_res.input.file.name}"
+            )
+            return ConversionStatus.PARTIAL_SUCCESS
+        return ConversionStatus.SUCCESS
 
     @classmethod
     def get_default_options(cls) -> AsrPipelineOptions:

tests/test_asr_pipeline.py

@@ -57,3 +57,29 @@ def test_asr_pipeline_conversion(test_audio_path):
     print(f"Transcribed text from {test_audio_path.name}:")
     for i, text_item in enumerate(texts):
         print(f"  {i + 1}: {text_item.text}")
+
+
+@pytest.fixture
+def silent_audio_path():
+    """Fixture to provide the path to a silent audio file."""
+    path = Path("./tests/data/audio/silent_1s.wav")
+    if not path.exists():
+        pytest.skip("Silent audio file for testing not found at " + str(path))
+    return path
+
+
+def test_asr_pipeline_with_silent_audio(silent_audio_path):
+    """
+    Test that the ASR pipeline correctly handles silent audio files
+    by returning a PARTIAL_SUCCESS status.
+    """
+    converter = get_asr_converter()
+    doc_result: ConversionResult = converter.convert(silent_audio_path)
+
+    # This test will FAIL initially, which is what we want.
+    assert doc_result.status == ConversionStatus.PARTIAL_SUCCESS, (
+        f"Status should be PARTIAL_SUCCESS for silent audio, but got {doc_result.status}"
+    )
+    assert len(doc_result.document.texts) == 0, (
+        "Document should contain zero text items"
+    )