fix(load): clear FileNotFoundError for missing audio path

[LauraGPT]

Problem (new-user UX)

Following the README quickstart with the placeholder path input="meeting.wav" (a file new users don't have) does not give a clear error. Instead it crashes deep inside the VAD with a cryptic:

TypeError: expected Tensor as element 1 in argument 0, but got str
  ... fsmn_vad_streaming/model.py: torch.cat((cache["prev_samples"], audio_sample_list[0]))

Root cause: in load_audio_text_image_video, audio is only loaded if os.path.exists(path). A non-existent path string falls through the gate and is passed downstream as a raw str.

Fix

Raise a clear FileNotFoundError naming the missing path, before the load gate. Scoped to data_type="sound" so text/PUNC inputs are unaffected.

Tested (fresh venv, as a new user)

CASE 1 missing file -> FileNotFoundError: Audio file not found: 'meeting.wav'. Pass a valid local file path, URL, numpy array, torch.Tensor, or bytes.
CASE 2 valid file   -> unchanged, transcribes normally

Found by walking through the README quickstart as a first-time user.

[gemini-code-assist]

Code Review

This pull request introduces a fail-fast check in 'load_audio_text_image_video' to raise a clear 'FileNotFoundError' when a local audio file path does not exist, preventing cryptic downstream errors. The review feedback correctly points out that this check should also apply when 'data_type' is None (which defaults to 'sound'), as the current implementation only triggers when 'data_type' is explicitly set to 'sound'.

Important

The consumer version of Gemini Code Assist on GitHub is being sunset. Starting June 18, 2026, new organization installations will be blocked, and all code review activity will officially cease on July 17, 2026.
For more details on the timeline and next steps, please review the Help Documentation.

[gemini-code-assist]

The check currently only validates when data_type == "sound". However, on line 114, data_type can also be None (which defaults to "sound" behavior). If data_type is None and the file does not exist, the check will be bypassed, and the function will still fail downstream with a cryptic error.\n\nWe should update the condition to also cover data_type is None.

    if (\n        isinstance(data_or_path_or_list, str)\n        and (data_type == "sound" or data_type is None)\n        and not data_or_path_or_list.startswith(("http://", "https://"))\n        and not os.path.exists(data_or_path_or_list)\n    ):

[LauraGPT]

Thanks! Addressed — the check now also covers data_type is None (which defaults to sound): data_type in (None, "sound").