"""Shared fixtures.Every test runs on CPU in float32. The interesting numerical properties here (logitparity between prefill and cached decode, bias injection actually changing the output) holdat any precision, and testing in float32 means a failure means a real bug rather than anaccumulated rounding difference."""from __future__ import annotationsimport pytestimport torchfrom elafry.config import PRESETS, AffectConfigfrom elafry.models.elafry import Elafrydef make_tiny_model(seed: int = 0, **affect_kwargs) -> Elafry: """A real Elafry, small enough that the whole suite runs in seconds.""" torch.manual_seed(seed) preset = PRESETS["elafry-tiny"] affect = AffectConfig(**affect_kwargs) return Elafry(preset["model"], affect)@pytest.fixturedef tiny_model() -> Elafry: return make_tiny_model(seed=0)@pytest.fixturedef tiny_ids() -> torch.Tensor: torch.manual_seed(1234) return torch.randint(0, PRESETS["elafry-tiny"]["model"].vocab_size, (2, 12))@pytest.fixture(autouse=True)def deterministic(): """Pin the seed around every test so a failure is reproducible.""" torch.manual_seed(0) yield