"""Publish the pod's own status and results to the Hub.

The control channel of last resort. RunPod's SSH proxy has been dropping roughly
two connections in three, corrupting long command lines, and requiring a PTY that
echoes everything back -- so every measurement taken through it has been lossy or
lost. The Hub has not failed once, and the pod already holds a token for it.

  push.py status <kept> <running>   one small JSON, overwritten each tick
  push.py done   <kept>             status plus the artifacts worth keeping
"""

import json
import os
import pathlib
import sys
import time

from huggingface_hub import HfApi

REPO = "btsee/oron-tts"
api = HfApi(token=os.environ["HF_TOKEN"])


def put(local: pathlib.Path, remote: str, msg: str) -> None:
    for attempt in range(3):
        try:
            api.upload_file(path_or_fileobj=str(local), path_in_repo=remote,
                            repo_id=REPO, commit_message=msg)
            return
        except Exception:
            time.sleep(2 * (attempt + 1))


def main() -> None:
    mode = sys.argv[1]
    kept = int(sys.argv[2]) if len(sys.argv) > 2 else 0
    running = sys.argv[3] if len(sys.argv) > 3 else "?"

    started = pathlib.Path("/workspace/.started_at")
    if not started.is_file():
        started.write_text(str(int(time.time())))
    elapsed = int(time.time()) - int(started.read_text().strip())

    status = {"mode": mode, "kept": kept, "cleaner_running": running,
              "elapsed_s": elapsed,
              "kept_per_min": round(kept * 60.0 / elapsed, 2) if elapsed else None,
              "utc": time.strftime("%Y-%m-%dT%H:%M:%SZ", time.gmtime())}
    p = pathlib.Path("/workspace/status.json")
    p.write_text(json.dumps(status, indent=2))
    put(p, "eval/v4_status.json", f"v4 status: {kept} clips kept")

    if mode != "done":
        return

    # Everything a decision could need, small enough to read from the Hub.
    for local, remote in (
        ("/workspace/bw_summary.json", "eval/v4_bw_summary.json"),
        ("/workspace/cv/eval_sentences.txt", "eval/v4_cv_eval_sentences.txt"),
        # write_bundle emits voices.json, not selection.json.
        ("/workspace/voices_v4/voices.json", "eval/v4_voices.json"),
        ("/workspace/cv.log", "eval/v4_cv.log"),
    ):
        lp = pathlib.Path(local)
        if lp.is_file():
            put(lp, remote, "v4 run artifacts")

    # The chosen prompts themselves: this is the deliverable.
    vd = pathlib.Path("/workspace/voices_v4")
    if vd.is_dir():
        for f in sorted(vd.iterdir()):
            if f.suffix in (".wav", ".txt", ".json"):
                put(f, f"eval/v4_voices/{f.name}", "v4 candidate reference prompts")

    # The manifest is the evidence behind the ranking; gzip keeps it small.
    man = pathlib.Path("/workspace/cv/manifest.jsonl")
    if man.is_file():
        import gzip
        gz = pathlib.Path("/workspace/manifest_v4.jsonl.gz")
        gz.write_bytes(gzip.compress(man.read_bytes()))
        put(gz, "eval/v4_manifest.jsonl.gz", "v4 manifest")


if __name__ == "__main__":
    main()