[Fix] Fix Watsonx Audio Transcription API (#17326)

* """
add

* fix transform_audio_transcription_request

* fix tests

* test_watsonx_transcription_request_body

litellm/llms/watsonx/audio_transcription/transformation.py

@@ -4,11 +4,17 @@ Translates from OpenAI's `/v1/audio/transcriptions` to IBM WatsonX's `/ml/v1/aud
 WatsonX follows the OpenAI spec for audio transcription.
 """
 
-from typing import List, Optional
+from typing import Any, Dict, List, Optional
 
 import litellm
+from litellm.litellm_core_utils.audio_utils.utils import process_audio_file
 from litellm.types.llms.openai import OpenAIAudioTranscriptionOptionalParams
+from litellm.types.llms.watsonx import WatsonXAudioTranscriptionRequestBody
+from litellm.types.utils import FileTypes
 
+from ...base_llm.audio_transcription.transformation import (
+    AudioTranscriptionRequestData,
+)
 from ...openai.transcriptions.whisper_transformation import (
     OpenAIWhisperAudioTranscriptionConfig,
 )
@@ -40,6 +46,60 @@ class IBMWatsonXAudioTranscriptionConfig(
             "timestamp_granularities",
         ]
 
+    def transform_audio_transcription_request(
+        self,
+        model: str,
+        audio_file: FileTypes,
+        optional_params: dict,
+        litellm_params: dict,
+    ) -> AudioTranscriptionRequestData:
+        """
+        Transform the audio transcription request for WatsonX.
+        
+        WatsonX expects multipart/form-data with:
+        - file: the audio file
+        - model: the model name (without watsonx/ prefix)
+        - project_id: the project ID (as form field, not query param)
+        - other optional params
+        """
+        # Use common utility to process the audio file
+        processed_audio = process_audio_file(audio_file)
+        
+        # Get API params to extract project_id
+        api_params = _get_api_params(params=optional_params.copy())
+        
+        # Initialize form data with required fields
+        form_data: WatsonXAudioTranscriptionRequestBody = {
+            "model": model,
+            "project_id": api_params.get("project_id", ""),
+        }
+        
+        # Add supported OpenAI params to form data
+        supported_params = self.get_supported_openai_params(model)
+        for key, value in optional_params.items():
+            if key in supported_params and value is not None:
+                form_data[key] = value  # type: ignore
+        
+        # Set default response_format for cost calculation
+        if "response_format" not in form_data or (
+            form_data.get("response_format") in ["text", "json"]
+        ):
+            form_data["response_format"] = "verbose_json"
+        
+        # Prepare files dict with the audio file
+        files = {
+            "file": (
+                processed_audio.filename,
+                processed_audio.file_content,
+                processed_audio.content_type,
+            )
+        }
+        
+        # Convert TypedDict to regular dict for AudioTranscriptionRequestData
+        form_data_dict: Dict[str, Any] = dict(form_data)
+        
+        return AudioTranscriptionRequestData(data=form_data_dict, files=files)
+
     def get_complete_url(
         self,
         api_base: Optional[str],
@@ -52,7 +112,9 @@ class IBMWatsonXAudioTranscriptionConfig(
         """
         Construct the complete URL for WatsonX audio transcription.
 
-        URL format: {api_base}/ml/v1/audio/transcriptions?version={version}&project_id={project_id}
+        URL format: {api_base}/ml/v1/audio/transcriptions?version={version}
+        
+        Note: project_id is sent as form data, not as a query parameter
         """
         # Get base URL
         url = self._get_base_url(api_base=api_base)
@@ -61,18 +123,10 @@ class IBMWatsonXAudioTranscriptionConfig(
         # Add the audio transcription endpoint
         url = f"{url}/ml/v1/audio/transcriptions"
 
-        # Get API params for project_id
-        api_params = _get_api_params(params=optional_params.copy())
-
-        # Add version parameter
-        api_version = optional_params.pop(
+        # Add version parameter (only version in query string, not project_id)
+        api_version = optional_params.get(
             "api_version", None
         ) or litellm.WATSONX_DEFAULT_API_VERSION
         url = f"{url}?version={api_version}"
 
-        # Add project_id parameter
-        project_id = api_params.get("project_id")
-        if project_id:
-            url = f"{url}&project_id={project_id}"
-
         return url

litellm/types/llms/watsonx.py

@@ -1,9 +1,7 @@
-import json
 from enum import Enum
-from typing import Any, List, Optional, Union
+from typing import List, Optional
 
-from pydantic import BaseModel
-from typing_extensions import TypedDict
+from typing_extensions import NotRequired, TypedDict
 
 
 class WatsonXAPIParams(TypedDict):
@@ -18,6 +16,36 @@ class WatsonXCredentials(TypedDict):
     token: Optional[str]
 
 
+class WatsonXAudioTranscriptionRequestBody(TypedDict):
+    """
+    WatsonX Audio Transcription API request body.
+    
+    Follows multipart/form-data format for WatsonX Whisper models.
+    See: https://cloud.ibm.com/apidocs/watsonx-ai
+    """
+
+    model: str
+    """Model name (e.g., 'whisper-large-v3-turbo')"""
+
+    project_id: str
+    """WatsonX project ID (required)"""
+
+    language: NotRequired[str]
+    """Language code (e.g., 'en', 'es')"""
+
+    prompt: NotRequired[str]
+    """Optional prompt to guide transcription"""
+
+    response_format: NotRequired[str]
+    """Response format: 'json', 'text', 'srt', 'verbose_json', 'vtt'"""
+
+    temperature: NotRequired[float]
+    """Sampling temperature (0-1)"""
+
+    timestamp_granularities: NotRequired[List[str]]
+    """Timestamp granularities: ['word', 'segment']"""
+
+
 class WatsonXAIEndpoint(str, Enum):
     TEXT_GENERATION = "/ml/v1/text/generation"
     TEXT_GENERATION_STREAM = "/ml/v1/text/generation_stream"

tests/test_litellm/llms/watsonx/audio_transcription/test_watsonx_audio_transcription_transformation.py

@@ -4,6 +4,7 @@ Tests for IBM WatsonX Audio Transcription.
 Validates that litellm.transcription transforms requests correctly for WatsonX.
 """
 
+import json
 import os
 import sys
 from unittest.mock import AsyncMock, MagicMock, patch
@@ -29,6 +30,7 @@ class TestWatsonXAudioTranscription:
             captured_request["url"] = str(kwargs.get("url", args[0] if args else None))
             captured_request["headers"] = kwargs.get("headers", {})
             captured_request["data"] = kwargs.get("data", {})
+            captured_request["files"] = kwargs.get("files", {})
             
             mock_response = MagicMock()
             mock_response.json.return_value = {
@@ -54,16 +56,30 @@ class TestWatsonXAudioTranscription:
         # Validate URL contains WatsonX audio transcription endpoint
         assert "/ml/v1/audio/transcriptions" in captured_request["url"]
         assert "version=" in captured_request["url"]
-        assert "project_id=test-project-123" in captured_request["url"]
+        # project_id should NOT be in URL (it should be in form data instead)
+        assert "project_id=test-project-123" not in captured_request["url"]
 
         # Validate headers contain WatsonX auth
         assert "Authorization" in captured_request["headers"]
         assert "Bearer test-bearer-token" in captured_request["headers"]["Authorization"]
+        
+        # Validate project_id is in form data, not URL
+        assert captured_request["data"].get("project_id") == "test-project-123"
+        
+        # Validate file is in files dict
+        assert "file" in captured_request["files"]
 
     @pytest.mark.asyncio
     async def test_watsonx_transcription_request_body(self):
         """
         Test that litellm.transcription sends correct request body for WatsonX.
+        
+        Validates that:
+        - Request uses multipart/form-data (data + files)
+        - Model name has watsonx/ prefix removed
+        - project_id is in form data, not URL
+        - Audio file is in files dict
+        - OpenAI params are included in form data
         """
         captured_request = {}
 
@@ -94,9 +110,24 @@ class TestWatsonXAudioTranscription:
             except Exception:
                 pass  # We just want to capture the request
 
-        # Validate request body contains expected fields
+        # Validate form data contains expected fields
         data = captured_request.get("data", {})
+
+        print("JSON DUMPS captured_request:")
+        print(json.dumps(captured_request, indent=4, default=str))
+        
+        # Model name should NOT have watsonx/ prefix
         assert data.get("model") == "whisper-large-v3-turbo"
+        
+        # project_id should be in form data
+        assert data.get("project_id") == "test-project-123"
+        
+        # OpenAI params should be in form data
         assert data.get("language") == "en"
         assert data.get("temperature") == 0.5
         assert data.get("response_format") == "verbose_json"  # Default for cost calculation
+        
+        # Validate file is in files dict (multipart/form-data)
+        files = captured_request.get("files", {})
+        assert "file" in files
+        assert isinstance(files["file"], tuple)  # Should be (filename, content, content_type)