import json
import wave
from pathlib import Path

import numpy as np

from app.audio import build_audio_profile, classify_audio, classify_sequence
from app.config import settings
from app.profile_store import Profile


def _write_tone(path: Path, frequency: float, duration: float = 1.2) -> None:
    sample_rate = 16000
    time = np.arange(int(sample_rate * duration)) / sample_rate
    envelope = np.minimum(1.0, time * 8) * np.minimum(1.0, (duration - time) * 8)
    samples = 0.7 * np.sin(2 * np.pi * frequency * time) * envelope
    pcm = (samples * 32767).astype("<i2")
    path.parent.mkdir(parents=True, exist_ok=True)
    with wave.open(str(path), "wb") as output:
        output.setnchannels(1)
        output.setsampwidth(2)
        output.setframerate(sample_rate)
        output.writeframes(pcm.tobytes())


def test_audio_profile_classifies_and_rejects(tmp_path: Path) -> None:
    profiles = tmp_path / "profiles"
    artifacts = tmp_path / "artifacts"
    settings.profiles_dir = profiles
    settings.artifacts_dir = artifacts
    directory = profiles / "test" / "tones" / "v1"
    config = {
        "code": "test.tones.v1",
        "analysisType": "AUDIO_CLASSIFICATION",
        "decision": {
            "maxDistance": 0.45,
            "minDistanceMargin": 0.02,
            "unknownLabel": "UNKNOWN",
        },
    }
    directory.mkdir(parents=True)
    (directory / "profile.json").write_text(json.dumps(config), encoding="utf-8")
    reference_tones = {
        "POWER_ON": 440,
        "POWER_OFF": 660,
        "ARMED": 880,
        "DISARMED": 1100,
        "FIND_VEHICLE_HORN": 1320,
    }
    for label, frequency in reference_tones.items():
        _write_tone(directory / "references" / label / f"{label.lower()}.wav", frequency)
    profile = Profile("test.tones.v1", "AUDIO_CLASSIFICATION", directory, config)

    build_audio_profile(profile)
    classified_paths = {}
    for label, frequency in reference_tones.items():
        path = tmp_path / f"{label.lower()}-test.wav"
        _write_tone(path, frequency)
        classified_paths[label] = path
    other_path = tmp_path / "other-test.wav"
    _write_tone(other_path, 1600)

    for label, path in classified_paths.items():
        assert classify_audio(profile, path)["label"] == label
    assert classify_audio(profile, other_path)["label"] == "UNKNOWN"


def test_audio_profile_aggregates_multiple_references(monkeypatch) -> None:
    config = {
        "decision": {
            "maxDistance": 0.45,
            "minDistanceMargin": 0.001,
            "topK": 3,
            "nearestWeight": 0.6,
            "unknownLabel": "UNKNOWN",
        }
    }
    profile = Profile("test.aggregate.v1", "AUDIO_CLASSIFICATION", Path("."), config)
    references = [np.array([[value]], dtype=np.float32) for value in range(1, 7)]
    distances = {1: 0.10, 2: 0.11, 3: 0.12, 4: 0.0, 5: 0.40, 6: 0.40}
    monkeypatch.setattr(
        "app.audio.sequence_distance",
        lambda _query, reference: distances[int(reference[0, 0])],
    )

    result = classify_sequence(
        profile,
        np.array([[0.0]], dtype=np.float32),
        references,
        np.array(["EXPECTED"] * 3 + ["OUTLIER"] * 3),
        np.array([f"sample-{index}.wav" for index in range(6)]),
    )

    assert result["label"] == "EXPECTED"
    assert result["evidence"]["neighborCount"] == 3
    assert result["thresholds"]["topK"] == 3


def test_audio_profile_applies_label_specific_threshold(monkeypatch) -> None:
    config = {
        "decision": {
            "maxDistance": 0.45,
            "minDistanceMargin": 0.015,
            "labelMaxDistance": {"FIND_VEHICLE_HORN": 0.35},
            "unknownLabel": "UNKNOWN",
        }
    }
    profile = Profile("test.threshold.v1", "AUDIO_CLASSIFICATION", Path("."), config)
    references = [np.array([[1.0]], dtype=np.float32), np.array([[2.0]], dtype=np.float32)]
    monkeypatch.setattr(
        "app.audio.sequence_distance",
        lambda _query, reference: 0.36 if int(reference[0, 0]) == 1 else 0.44,
    )

    result = classify_sequence(
        profile,
        np.array([[0.0]], dtype=np.float32),
        references,
        np.array(["FIND_VEHICLE_HORN", "POWER_OFF"]),
        np.array(["horn.wav", "power-off.wav"]),
    )

    assert result["label"] == "UNKNOWN"
    assert result["thresholds"]["maxDistance"] == 0.35