diff --git a/tests/unit/llms/gemini/audio_transcription/test_gemini_audio_transcription_transformation.py b/tests/unit/llms/gemini/audio_transcription/test_gemini_audio_transcription_transformation.py index ab2ee82b604..de3a1653e6b 100644 --- a/tests/unit/llms/gemini/audio_transcription/test_gemini_audio_transcription_transformation.py +++ b/tests/unit/llms/gemini/audio_transcription/test_gemini_audio_transcription_transformation.py @@ -220,6 +220,22 @@ class TestTransformRequest: "diarization_mode": "speaker", } + def test_global_drop_params_removes_keywords_with_word_timestamps(self, config, monkeypatch): + monkeypatch.setattr(litellm, "drop_params", True) + + request_data = config.transform_audio_transcription_request( + model="gemini-3.5-transcribe", + audio_file=("sample.wav", AUDIO_BYTES, "audio/wav"), + optional_params={ + "keywords": ["alpha"], + "timestamp_granularities": ["word"], + }, + litellm_params={}, + ) + transcription_config = request_data.data["generation_config"]["transcription_config"] + assert "custom_vocabulary" not in transcription_config + assert transcription_config["mode"]["timestamp_granularities"] == ("word",) + def test_word_timestamp_granularity_maps_to_verbatim_diarization_mode(self, config): request_data = config.transform_audio_transcription_request( model="gemini-3.5-transcribe", @@ -253,6 +269,18 @@ class TestTransformRequest: } } + @pytest.mark.parametrize("response_format", ["srt", "vtt"]) + def test_keywords_with_subtitle_response_format_raise(self, config, response_format, monkeypatch): + monkeypatch.setattr(litellm, "drop_params", False) + + with pytest.raises(litellm.UnsupportedParamsError, match="custom vocabulary with word timestamps"): + config.transform_audio_transcription_request( + model="gemini-3.5-transcribe", + audio_file=("sample.wav", AUDIO_BYTES, "audio/wav"), + optional_params={"keywords": ["alpha"], "response_format": response_format}, + litellm_params={}, + ) + @pytest.mark.parametrize("response_format", ["json", "text", "verbose_json"]) def test_non_subtitle_response_format_sends_no_mode(self, config, response_format): request_data = config.transform_audio_transcription_request(