Next-Echo / inference.py
Lamapi's picture
Publish Next-Echo model
e187f2f verified
Raw History Blame Contribute Delete
2.3 kB
"""Next Echo inference (NumPy only): find occurrences of an example sound in a recording.
Usage:
python inference.py --query clap.wav --recording street.wav
python inference.py --selftest
Audio must be 16 kHz mono 16-bit WAV for the CLI (resample elsewhere if needed).
"""
import argparse
import json
import os
import sys
import wave
HERE = os.path.dirname(os.path.abspath(__file__))
sys.path.insert(0, HERE)
import numpy as np # noqa: E402
import echo_core as ec # noqa: E402
from nlrt import load_model as _load # noqa: E402
def load(path=os.path.join(HERE, "model.json")):
return ec.load_model(_load(path))
def read_wav(path):
with wave.open(path, "rb") as w:
assert w.getframerate() == 16000 and w.getnchannels() == 1 and w.getsampwidth() == 2, "need 16 kHz mono 16-bit"
return np.frombuffer(w.readframes(w.getnframes()), dtype="<i2") / 32768.0
def find(model, query, recording, method="learned"):
times, learned, base = ec.similarity_curves(model, query, recording)
nxt, bl = ec.final_scores(learned, base, model["score"])
sims = nxt if method == "learned" else bl
th = model["thresholds"]["learned" if method == "learned" else "baseline"]
return [{"start": round(a, 3), "end": round(b, 3), "score": round(s, 4)} for a, b, s in ec.detect(times, sims, th)]
def _selftest():
model = load()
with open(os.path.join(HERE, "selftest.json"), encoding="utf-8") as f:
t = json.load(f)
q, r = np.asarray(t["query"]) / 32768.0, np.asarray(t["recording"]) / 32768.0
_, learned, base = ec.similarity_curves(model, q, r)
diff = max(float(np.abs(learned - np.asarray(t["learned"])).max()), float(np.abs(base - np.asarray(t["base"])).max()))
ok = diff < 1e-6
print(json.dumps({"passed": ok, "max_abs_diff": diff, "windows": len(learned)}))
return ok
if __name__ == "__main__":
ap = argparse.ArgumentParser()
ap.add_argument("--selftest", action="store_true")
ap.add_argument("--query")
ap.add_argument("--recording")
ap.add_argument("--method", default="learned", choices=["learned", "logmel"])
a = ap.parse_args()
if a.selftest:
sys.exit(0 if _selftest() else 1)
print(json.dumps(find(load(), read_wav(a.query), read_wav(a.recording), a.method), indent=1))