#!/usr/bin/env python3
"""Create synthetic-only speech fixtures. Never accepts a recording or sends data to a provider.

The random canaries and expected facts live in the local manifest only; the Dify
probe sends the audio, schema and recording time but never this ground truth.
Requires macOS say plus ffmpeg/ffprobe. Refuses to replace an existing manifest.
"""
import argparse
import hashlib
import json
import pathlib
import secrets
import subprocess
import tempfile


def run(command):
    return subprocess.run(command, check=True, capture_output=True, text=True).stdout


def duration(path):
    return float(run(["ffprobe", "-v", "error", "-show_entries", "format=duration", "-of", "default=nw=1:nk=1", str(path)]))


def main():
    parser = argparse.ArgumentParser()
    parser.add_argument("--output", type=pathlib.Path, required=True)
    parser.add_argument("--voice", default="Tingting")
    args = parser.parse_args()
    output = args.output.resolve()
    if (output / "manifest.json").exists() or any((output / (case + ".mp3")).exists() for case in ("short", "medium", "long")):
        raise SystemExit("Fixture identity already exists; reuse it instead of regenerating or resubmitting.")
    output.mkdir(parents=True, exist_ok=True)
    output.chmod(0o700)
    manifest = {"generator": "followup-audio-synthetic-v1", "synthetic": True,
                "recorded_at": "2026-09-29 10:00:00", "fixtures": []}
    words = ["紫色海豚", "水晶风车", "绿色蜻蜓", "银色树叶", "蓝色鲸鱼", "金色河流", "红色石榴", "白色云朵"]
    with tempfile.TemporaryDirectory(prefix="followup-audio-synthetic-") as temporary:
        temp = pathlib.Path(temporary)
        middle_text = "这是纯合成的声音长度测试，不对应任何真实患者。现在继续校验这段音频的可听性。重复的测试语句不是医疗事实，不应反复生成随访记录。"
        run(["say", "-v", args.voice, "-r", "190", "-o", str(temp / "middle.aiff"), middle_text])
        for case, seconds, systolic, diastolic in (("short", 60, 131, 86), ("medium", 900, 139, 88), ("long", 3598, 147, 91)):
            head_canary = "".join(secrets.SystemRandom().sample(words, 3))
            tail_canary = "".join(secrets.SystemRandom().sample(words, 3))
            while tail_canary == head_canary:
                tail_canary = "".join(secrets.SystemRandom().sample(words, 3))
            head = f"这是一段人工合成的音频测试。开始校验词是，{head_canary}。昨天晚上九点，收缩压一百二十六，舒张压八十二。前天下午两点，餐后血糖七点二。"
            tail = f"现在是这段音频的最后部分。今天早晨七点，收缩压{systolic}，舒张压{diastolic}。末尾校验词是，{tail_canary}。合成测试到此结束。"
            run(["say", "-v", args.voice, "-r", "190", "-o", str(temp / "head.aiff"), head])
            run(["say", "-v", args.voice, "-r", "190", "-o", str(temp / "tail.aiff"), tail])
            filler_seconds = seconds - duration(temp / "head.aiff") - duration(temp / "tail.aiff")
            if filler_seconds <= 0:
                raise SystemExit("Speech is longer than short fixture; increase speaking rate or duration")
            path = output / (case + ".mp3")
            filters = f"[0:a]aresample=16000,aformat=channel_layouts=mono[h];[1:a]aresample=16000,aformat=channel_layouts=mono,atrim=duration={filler_seconds},asetpts=PTS-STARTPTS[m];[2:a]aresample=16000,aformat=channel_layouts=mono[t];[h][m][t]concat=n=3:v=0:a=1[out]"
            run(["ffmpeg", "-nostdin", "-hide_banner", "-loglevel", "error", "-i", str(temp / "head.aiff"),
                 "-stream_loop", "-1", "-i", str(temp / "middle.aiff"), "-i", str(temp / "tail.aiff"),
                 "-filter_complex", filters, "-map", "[out]", "-t", str(seconds), "-codec:a", "libmp3lame", "-b:a", "32k", "-y", str(path)])
            path.chmod(0o600)
            actual = duration(path)
            sha = hashlib.file_digest(path.open("rb"), "sha256").hexdigest()
            manifest["fixtures"].append({"case": case, "file": path.name, "sha256": sha, "duration_seconds": actual,
                "synthetic": True, "generator": manifest["generator"], "expected": {"canaries": [head_canary, tail_canary], "facts": [
                    {"kind": "blood", "record_date": "2026-09-28", "record_time": "21:00", "values": {"systolic_pressure": 126, "diastolic_pressure": 82}},
                    {"kind": "blood", "record_date": "2026-09-27", "record_time": "14:00", "values": {"postprandial_blood_sugar": 7.2}},
                    {"kind": "blood", "record_date": "2026-09-29", "record_time": "07:00", "values": {"systolic_pressure": systolic, "diastolic_pressure": diastolic}},
                ]}})
            print(json.dumps({"case": case, "path": str(path), "duration_seconds": actual, "sha256": sha, "synthetic": True}))
    manifest_path = output / "manifest.json"
    manifest_path.write_text(json.dumps(manifest, ensure_ascii=False, indent=2) + "\n")
    manifest_path.chmod(0o600)
    reopened = json.loads(manifest_path.read_text())
    assert len(reopened["fixtures"]) == 3
    print(json.dumps({"manifest": str(manifest_path), "synthetic": True, "fixtures": 3}))


if __name__ == "__main__":
    main()