Palash DebnathandClaude Opus 4.8
c6a55794da
Surfaces a routing verdict for the CURRENTLY-SELECTED TTS engine in the two system-health surfaces, so a CPU fallback / unavailable-GPU is heard about before a slow or failed synth — the no-silent-fallback contract, read-only. - `tts_backend.active_routing()` + `gpu_routing_verdict()`: the active engine's routing derived from list_backends() (byte-identical to the matrix) plus the host compute summary (family + VRAM from the canonical probe). Never raise. - `/system/diagnose` gains a `gpu_routing` check: accelerated→ok, accelerated-with-caveat / cpu_fallback→warn (+ actionable hint), cpu_only→ok (no-GPU host is the expected normal state — never noise-warns), unavailable→ fail, no-engine→warn. ASCII-safe detail strings (the text dump enforces ASCII). - `/setup/preflight` gains an "Active engine routing" check + an explicit `gpu_routing` object on PreflightResponse (a real field — the response has no extra="allow", so it would otherwise be dropped). `device` gains `gpu_family` (ROCm-vs-CUDA aware) + `vram_gb`. New `GpuRouting` schema. Tests: gpu_routing_verdict (host + active-engine + degraded), diagnose status mapping across all 6 states + never-raises, preflight gpu_routing object + check + device.gpu_family. Existing diagnose/preflight tests stay green (checks are additive; the report's top-level key set is unchanged). Deferred (documented): synth-time routing headers/WS-frames at the 3 synth entry points. Selection is already hard-gated (PR 3 select_engine), and the matrix (PR 5) + this preflight/diagnose verdict surface the situation — the synth-time signal is incremental belt-and-suspenders for the env-var-pinned edge and is best validated interactively. Tracked as a #21 follow-up. Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
87 lines
3.8 KiB
Python
87 lines
3.8 KiB
Python
"""Active-engine GPU routing verdict (#21 PR 4) — the no-silent-fallback
|
|
surface for preflight + diagnose.
|
|
|
|
Covers `tts_backend.gpu_routing_verdict()` (built from active_routing + the host
|
|
probe) and `diagnose._check_gpu_routing()`'s status mapping. Both are exercised
|
|
with mocks so no torch / full app import is needed.
|
|
"""
|
|
from __future__ import annotations
|
|
|
|
import pytest
|
|
|
|
from core.device_caps import HostCaps
|
|
from core.diagnose import OK, WARN, FAIL
|
|
|
|
|
|
# ── gpu_routing_verdict ─────────────────────────────────────────────────────
|
|
|
|
def test_verdict_uses_active_routing_and_host(monkeypatch):
|
|
from services import tts_backend as tb
|
|
monkeypatch.setattr(tb, "active_routing", lambda: {
|
|
"engine": "omnivoice", "available": True,
|
|
"effective_device": "cpu", "routing_status": "cpu_fallback",
|
|
"routing_reason": "engine has no CUDA path; running on CPU",
|
|
})
|
|
monkeypatch.setattr(
|
|
"core.device_caps.detect_host_caps",
|
|
lambda: HostCaps(family="cuda", available_families=("cuda", "cpu"), vram_gb=24.0),
|
|
)
|
|
v = tb.gpu_routing_verdict()
|
|
assert v["engine"] == "omnivoice"
|
|
assert v["routing_status"] == "cpu_fallback"
|
|
assert v["host_family"] == "cuda"
|
|
assert v["vram_gb"] == 24.0
|
|
|
|
|
|
def test_verdict_degrades_when_no_active_engine(monkeypatch):
|
|
from services import tts_backend as tb
|
|
monkeypatch.setattr(tb, "active_routing", lambda: None)
|
|
monkeypatch.setattr(
|
|
"core.device_caps.detect_host_caps",
|
|
lambda: HostCaps(family="cpu", available_families=("cpu",)),
|
|
)
|
|
v = tb.gpu_routing_verdict()
|
|
assert v["engine"] is None
|
|
assert v["routing_status"] == "none"
|
|
assert v["host_family"] == "cpu"
|
|
|
|
|
|
# ── diagnose._check_gpu_routing status mapping ──────────────────────────────
|
|
|
|
@pytest.mark.parametrize("verdict,expected_status", [
|
|
({"engine": "e", "effective_device": "cuda", "routing_status": "accelerated",
|
|
"routing_reason": None, "host_family": "cuda", "vram_gb": 24.0}, OK),
|
|
({"engine": "e", "effective_device": "cuda", "routing_status": "accelerated",
|
|
"routing_reason": "CUDA selected, but: may fail at kernel launch",
|
|
"host_family": "cuda", "vram_gb": 24.0}, WARN),
|
|
({"engine": "e", "effective_device": "cpu", "routing_status": "cpu_fallback",
|
|
"routing_reason": "engine has no CUDA path; running on CPU",
|
|
"host_family": "cuda", "vram_gb": 24.0}, WARN),
|
|
({"engine": "e", "effective_device": "cpu", "routing_status": "cpu_only",
|
|
"routing_reason": None, "host_family": "cpu", "vram_gb": 0.0}, OK),
|
|
({"engine": "e", "effective_device": "cuda", "routing_status": "unavailable",
|
|
"routing_reason": "requires cuda; this host has cpu",
|
|
"host_family": "cpu", "vram_gb": 0.0}, FAIL),
|
|
({"engine": None, "effective_device": None, "routing_status": "none",
|
|
"routing_reason": None, "host_family": "cpu", "vram_gb": 0.0}, WARN),
|
|
])
|
|
def test_diagnose_routing_status_mapping(monkeypatch, verdict, expected_status):
|
|
import core.diagnose as diag
|
|
monkeypatch.setattr("services.tts_backend.gpu_routing_verdict", lambda: verdict)
|
|
check = diag._check_gpu_routing()
|
|
assert check["id"] == "gpu_routing"
|
|
assert check["status"] == expected_status
|
|
if expected_status in (WARN, FAIL):
|
|
assert check["hint"], "actionable hint required on warn/fail"
|
|
|
|
|
|
def test_diagnose_routing_never_raises(monkeypatch):
|
|
import core.diagnose as diag
|
|
|
|
def _boom():
|
|
raise RuntimeError("probe exploded")
|
|
|
|
monkeypatch.setattr("services.tts_backend.gpu_routing_verdict", _boom)
|
|
check = diag._check_gpu_routing()
|
|
assert check["status"] == WARN # degrades, doesn't propagate
|