community[patch]: Skip OpenAIWhisperParser extremely small audio ch…

…unks to avoid api error (langchain-ai#11450)

**Description**
This PR addresses a rare issue in `OpenAIWhisperParser` that causes it
to crash when processing an audio file with a duration very close to the
class's chunk size threshold of 20 minutes.

**Issue**
langchain-ai#11449

**Dependencies**
None

**Tag maintainer**
@agola11 @eyurtsev 

**Twitter handle**
leonardodiegues

---------

Co-authored-by: Leonardo Diegues <[email protected]>
Co-authored-by: Bagatur <[email protected]>

libs/community/langchain_community/document_loaders/parsers/audio.py

@@ -13,10 +13,22 @@
 
 class OpenAIWhisperParser(BaseBlobParser):
     """Transcribe and parse audio files.
-    Audio transcription is with OpenAI Whisper model."""
 
-    def __init__(self, api_key: Optional[str] = None):
+    Audio transcription is with OpenAI Whisper model.
+
+    Args:
+        api_key: OpenAI API key
+        chunk_duration_threshold: minimum duration of a chunk in seconds
+            NOTE: According to the OpenAI API, the chunk duration should be at least 0.1
+            seconds. If the chunk duration is less or equal than the threshold,
+            it will be skipped.
+    """
+
+    def __init__(
+        self, api_key: Optional[str] = None, *, chunk_duration_threshold: float = 0.1
+    ):
         self.api_key = api_key
+        self.chunk_duration_threshold = chunk_duration_threshold
 
     def lazy_parse(self, blob: Blob) -> Iterator[Document]:
         """Lazily parse the blob."""
@@ -57,6 +69,9 @@ def lazy_parse(self, blob: Blob) -> Iterator[Document]:
         for split_number, i in enumerate(range(0, len(audio), chunk_duration_ms)):
             # Audio chunk
             chunk = audio[i : i + chunk_duration_ms]
+            # Skip chunks that are too short to transcribe
+            if chunk.duration_seconds <= self.chunk_duration_threshold:
+                continue
             file_obj = io.BytesIO(chunk.export(format="mp3").read())
             if blob.source is not None:
                 file_obj.name = blob.source + f"_part_{split_number}.mp3"