"""Heli tekstiks Groq API-ga (Whisper large-v3, 99 keelt: it, ru, es, de, en, et …). Tasuta tase: 8 h heli/päev, 25 MB/fail.
Võti: SDR/secrets/groq_api_key.txt (ei lähe kunagi blogisse, skriptidesse ega logisse). Võtme saab console.groq.com → API Keys.
Kasutus: python transcribe_groq.py heli.wav [--lang it] [--model whisper-large-v3|whisper-large-v3-turbo] [--prompt "raadioamatöörid, kutsungid"]
         [--translate]  -> tõlgib inglise keelde (translations endpoint)
Väljund: tekst + <heli>.groq.json (verbose: keel, segmendid ajatemplitega).
"""
import argparse, json, os, sys, urllib.request, urllib.error, uuid
from pathlib import Path

ROOT = Path(__file__).resolve().parent.parent
API = "https://api.groq.com/openai/v1/audio/"


def api_key():
    f = ROOT / "secrets" / "groq_api_key.txt"
    if not f.exists():
        sys.exit("puudub secrets\\groq_api_key.txt – võta võti console.groq.com → API Keys ja salvesta sinna (üks rida)")
    return f.read_text(encoding="utf-8").strip()


def post(path, fields, endpoint):
    boundary = "----groq" + uuid.uuid4().hex
    body = b""
    for k, v in fields.items():
        body += f"--{boundary}\r\nContent-Disposition: form-data; name=\"{k}\"\r\n\r\n{v}\r\n".encode()
    body += (f"--{boundary}\r\nContent-Disposition: form-data; name=\"file\"; filename=\"{os.path.basename(path)}\"\r\n"
             f"Content-Type: audio/wav\r\n\r\n").encode() + open(path, "rb").read() + b"\r\n"
    body += f"--{boundary}--\r\n".encode()
    req = urllib.request.Request(API + endpoint, data=body, method="POST", headers={
        "Authorization": "Bearer " + api_key(), "Content-Type": f"multipart/form-data; boundary={boundary}",
        "User-Agent": "eeter-sdr/1.0"})
    try:
        with urllib.request.urlopen(req, timeout=120) as r:
            return json.loads(r.read().decode("utf-8"))
    except urllib.error.HTTPError as e:
        sys.exit(f"HTTP {e.code}: {e.read().decode('utf-8', 'replace')[:500]}")


def main():
    ap = argparse.ArgumentParser()
    ap.add_argument("wav"); ap.add_argument("--lang", default=None, help="ISO-639-1 (it, ru, es, de, en, et); puudub = tuvastab ise")
    ap.add_argument("--model", default="whisper-large-v3"); ap.add_argument("--prompt", default=None)
    ap.add_argument("--translate", action="store_true"); ap.add_argument("--out", default=None)
    a = ap.parse_args()
    size = os.path.getsize(a.wav)
    if size > 25 * 1024 * 1024:
        sys.exit(f"fail {size/1e6:.0f} MB > 25 MB (tasuta tase) – tee MP3 (to_flac.ps1 -Format mp3) või lõika lühemaks")
    fields = {"model": a.model, "response_format": "verbose_json", "temperature": "0"}
    if a.lang and not a.translate:
        fields["language"] = a.lang
    if a.prompt:
        fields["prompt"] = a.prompt
    r = post(a.wav, fields, "translations" if a.translate else "transcriptions")
    text = (r.get("text") or "").strip()
    lang = r.get("language", "?"); dur = r.get("duration", 0)
    print(f"[{lang}, {dur:.0f} s, {a.model}{', tõlge en' if a.translate else ''}]")
    print(text if text else "(tühi)")
    segs = r.get("segments") or []
    if segs:
        print("--- segmendid:")
        for s in segs:
            flag = "" if s.get("no_speech_prob", 0) < 0.5 else "  (kõne? %.0f%%)" % (100 * (1 - s["no_speech_prob"]))
            print(f"  {s['start']:6.1f}–{s['end']:6.1f}  {s['text'].strip()}{flag}")
    outj = Path(a.wav).with_suffix(".groq.json")
    outj.write_text(json.dumps(r, ensure_ascii=False, indent=1), encoding="utf-8")
    if a.out:
        Path(a.out).write_text(text + "\n", encoding="utf-8")


if __name__ == "__main__":
    main()
