slop rewriteHEADmain

1 files changed, 0 insertions, 70 deletions

diff --git a/lib/tests/conftest.py b/lib/tests/conftest.py
deleted file mode 100644
index 885d22b..0000000
--- a/lib/tests/conftest.py
+++ /dev/null
@@ -1,70 +0,0 @@
-from __future__ import annotations
-
-import sys
-from pathlib import Path
-
-import numpy as np
-import pytest
-
-SRC = Path(__file__).resolve().parents[1] / "src"
-sys.path.insert(0, str(SRC))
-sys.path.insert(0, str(Path(__file__).resolve().parent))
-
-SR = 44100
-
-
-def speechish(
-    dur: float,
-    sr: int = SR,
-    level_dbfs: float = -20.0,
-    seed: int = 0,
-) -> np.ndarray:
-    rng = np.random.default_rng(seed)
-    n = int(sr * dur)
-    t = np.arange(n) / sr
-    f0 = 110.0 * (1.0 + 0.02 * np.sin(2 * np.pi * 0.9 * t))
-    phase = 2 * np.pi * np.cumsum(f0) / sr
-    x = np.zeros(n)
-    for k in range(1, 9):
-        x += (1.0 / k**1.3) * np.sin(k * phase + 0.3 * k)
-    syll = 0.5 + 0.5 * np.sin(2 * np.pi * 3.0 * t + float(rng.uniform(0, 6)))
-    pauses = (np.sin(2 * np.pi * 0.5 * t) > -0.6).astype(float)
-    env = np.clip(syll, 0.02, 1.0) ** 0.6 * np.maximum(pauses, 0.05)
-    x = x * env
-    x /= np.max(np.abs(x)) + 1e-12
-    return (x * (10 ** (level_dbfs / 20.0))).astype(np.float32)
-
-
-def sine(freq: float, dur: float, sr: int = SR, peak_dbfs: float = -20.0) -> np.ndarray:
-    t = np.arange(int(sr * dur)) / sr
-    return (10 ** (peak_dbfs / 20.0) * np.sin(2 * np.pi * freq * t)).astype(np.float32)
-
-
-def band_db(x: np.ndarray, sr: int, lo: float, hi: float) -> float:
-    from scipy import signal
-
-    sos = signal.butter(4, [lo, hi], btype="bandpass", fs=sr, output="sos")
-    y = signal.sosfilt(sos, x.astype(np.float64))
-    r = np.sqrt(np.mean(np.square(y)))
-    if r <= 0:
-        return -120.0
-    return float(20 * np.log10(r))
-
-
-@pytest.fixture
-def sr() -> int:
-    return SR
-
-
-@pytest.fixture
-def speech() -> np.ndarray:
-    return speechish(6.0, level_dbfs=-20.0)
-
-
-@pytest.fixture
-def noisy_speech(sr, speech) -> np.ndarray:
-    rng = np.random.default_rng(7)
-    noise = rng.standard_normal(speech.size)
-    noise *= (10 ** (-48.0 / 20.0)) / np.sqrt(np.mean(np.square(noise)))
-    hum = 0.003 * np.sin(2 * np.pi * 50.0 * np.arange(speech.size) / sr)
-    return (speech + noise + hum).astype(np.float32)