Palash DebnathandClaude Opus 4.8
c3b2346759
Builds on the device probe from PR 1. Backend-only; the routing keys are wired into /engines in PR 3. - #390 closed: MLXAudioBackend / MLXWhisperBackend now call the shared `core.device_caps.mlx_supported()` gate FIRST, before importing the package. On Linux/Windows/mac-Intel they report unavailable and never advertise a usable `mps` route, even with a stray mlx wheel installed. Replaces the ASR backend's ad-hoc inline MPS check with the one shared rule. (The Wave-4.4 OSError/RuntimeError import-guard is preserved — it now lives behind the platform gate; its test forces the gate open so the guard stays the path under test.) - `ASRBackend` ABC gains `gpu_compat: tuple[str, ...] = ("cpu",)` mirroring TTSBackend, and each subclass declares its real targets: whisperx/faster-whisper → (cuda,cpu); mlx-whisper → (mps,cpu); pytorch-whisper → (cuda,mps,cpu); nemo/funasr → (cuda,cpu); moonshine → (cpu,). Inert until PR 3 serializes them. - IndexTTS2 declares `gpu_compat = ("cuda","cpu")` so it stops advertising the inherited CPU-only default. - ROCm is deliberately NOT claimed for any ASR engine (or for IndexTTS2): CTranslate2 has no upstream HIP build, and an unverified `rocm` claim would route ROCm hosts to a broken GPU path — strictly worse than the honest `cpu_fallback` the resolver already emits ("declares CUDA only; ROCm not in its compat set"). The per-engine TTS ROCm audit is a tracked follow-up that will verify each path before claiming it. Tests: MLX gate regression (both backends, on/off Apple), ASR gpu_compat tuples + no-false-rocm invariant, IndexTTS2 override; existing MLX import-guard test updated for the new gate ordering. Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
51 lines
2.0 KiB
Python
51 lines
2.0 KiB
Python
"""#390 regression: MLX-Audio and MLX-Whisper must report unavailable (and never
|
|
advertise a usable `mps` route) on non-Apple-Silicon hosts, even with the
|
|
package importable. Both backends now route through the shared
|
|
``core.device_caps.mlx_supported()`` gate BEFORE importing the package.
|
|
|
|
Backend classes are resolved at RUNTIME (other suites purge ``services.*`` from
|
|
``sys.modules``; importing the classes at module scope risks stale references
|
|
depending on test order).
|
|
"""
|
|
from __future__ import annotations
|
|
|
|
import pytest
|
|
|
|
_MLX_BACKENDS = [
|
|
("services.tts_backend", "MLXAudioBackend"),
|
|
("services.asr_backend", "MLXWhisperBackend"),
|
|
]
|
|
|
|
|
|
def _resolve(module_path, cls_name):
|
|
import importlib
|
|
return getattr(importlib.import_module(module_path), cls_name)
|
|
|
|
|
|
@pytest.mark.parametrize("module_path,cls_name", _MLX_BACKENDS)
|
|
def test_gate_blocks_when_mlx_unsupported(module_path, cls_name, monkeypatch):
|
|
monkeypatch.setattr(
|
|
"core.device_caps.mlx_supported",
|
|
lambda: (False, "MLX requires Apple Silicon; this host is linux/x86_64"),
|
|
)
|
|
ok, why = _resolve(module_path, cls_name).is_available()
|
|
assert ok is False
|
|
assert "Apple Silicon" in why
|
|
|
|
|
|
@pytest.mark.parametrize("module_path,cls_name", _MLX_BACKENDS)
|
|
def test_gate_passthrough_when_supported(module_path, cls_name, monkeypatch):
|
|
# When the platform gate passes, is_available proceeds to the package import.
|
|
monkeypatch.setattr("core.device_caps.mlx_supported", lambda: (True, ""))
|
|
ok, why = _resolve(module_path, cls_name).is_available()
|
|
# Proof the gate didn't short-circuit: control reached the package-import
|
|
# branch — so it's either genuinely available, or it reports the *package*
|
|
# error ("… unavailable"), never the platform-gate string.
|
|
assert ok or "unavailable" in why
|
|
assert not why.startswith("MLX requires Apple Silicon")
|
|
|
|
|
|
@pytest.mark.parametrize("module_path,cls_name", _MLX_BACKENDS)
|
|
def test_mlx_backends_declare_mps_cpu(module_path, cls_name):
|
|
assert _resolve(module_path, cls_name).gpu_compat == ("mps", "cpu")
|