| |
|
|
| import ast |
| import json |
| import pathlib |
|
|
| import pytest |
|
|
| ROOT = pathlib.Path(__file__).resolve().parent.parent |
| GEN = ROOT / "scripts" / "gen_fish_cot_samples.py" |
| FIT = ROOT / "scripts" / "fit_duration_model.py" |
|
|
|
|
| def _consts(src, fn): |
| tree = ast.parse(src) |
| for n in ast.walk(tree): |
| if isinstance(n, ast.FunctionDef) and n.name == fn: |
| return ast.unparse(n) |
| raise AssertionError(f"{fn} not found") |
|
|
|
|
| def test_predict_frames_matches_hand_computation(): |
| src = GEN.read_text(encoding="utf-8") |
| ns = {} |
| fn = _consts(src, "predict_frames") |
| exec(compile(fn, "<gen>", "exec"), ns) |
| predict = ns["predict_frames"] |
|
|
| dm = {"coef": [40.0, 2.0, 0.5, 7.0, -3.0]} |
| base = 40 + 2 * 3 + 0.5 * 13 |
| assert predict(dm, {"prompt": "one two three", "pace": "slow"}) \ |
| == pytest.approx(base + 7.0) |
| assert predict(dm, {"prompt": "one two three", "pace": "fast"}) \ |
| == pytest.approx(base - 3.0) |
| assert predict(dm, {"prompt": "one two three"}) == pytest.approx(base) |
| assert predict(dm, {"prompt": "one two three", "pace": "BRISK"}) \ |
| == pytest.approx(base) |
|
|
|
|
| def test_predict_frames_rejects_stale_model(): |
| src = GEN.read_text(encoding="utf-8") |
| ns = {} |
| exec(compile(_consts(src, "predict_frames"), "<gen>", "exec"), ns) |
| with pytest.raises(ValueError): |
| ns["predict_frames"]({"coef": [1.0, 2.0, 3.0]}, {"prompt": "a b"}) |
|
|
|
|
| def test_predict_frames_has_a_floor(): |
| src = GEN.read_text(encoding="utf-8") |
| ns = {} |
| exec(compile(_consts(src, "predict_frames"), "<gen>", "exec"), ns) |
| assert ns["predict_frames"]({"coef": [-1e6, 0, 0, 0, 0]}, |
| {"prompt": "a b c"}) >= 20.0 |
|
|
|
|
| def test_window_defaults_are_a_real_window(): |
| tree = ast.parse(GEN.read_text(encoding="utf-8")) |
| got = {} |
| for n in ast.walk(tree): |
| if (isinstance(n, ast.Call) |
| and getattr(n.func, "attr", "") == "add_argument" |
| and n.args and isinstance(n.args[0], ast.Constant) |
| and n.args[0].value in ("--dur_lo", "--dur_hi")): |
| got[n.args[0].value] = next( |
| k.value.value for k in n.keywords if k.arg == "default") |
| assert got["--dur_lo"] < 1.0 < got["--dur_hi"], got |
| assert got["--dur_lo"] > 0.0 |
|
|
|
|
| def test_duration_bounding_is_off_by_default(): |
| tree = ast.parse(GEN.read_text(encoding="utf-8")) |
| for n in ast.walk(tree): |
| if (isinstance(n, ast.Call) |
| and getattr(n.func, "attr", "") == "add_argument" |
| and n.args and isinstance(n.args[0], ast.Constant) |
| and n.args[0].value == "--duration_model"): |
| default = next((k.value.value for k in n.keywords |
| if k.arg == "default"), "MISSING") |
| assert default is None |
| return |
| raise AssertionError("--duration_model not found") |
|
|
|
|
| def test_max_frames_is_the_bounded_one(): |
| src = GEN.read_text(encoding="utf-8") |
| assert "max_frames=max_fr" in src |
| assert "max_frames=args.max_frames" not in src |
|
|
