#!/usr/bin/env python3
"""Re-runnable OpenAI model discovery: GET /v1/models -> sanitized snapshot + diff against the previous run.

- Calls GET /v1/models (free) via scripts/live.py (auth injected, call logged to reports/live-requests.jsonl).
- Diffs the returned ids against sources/openai/models-api-raw.json (previous run): prints NEW / REMOVED ids and
  any change of `shutdown_date`, and writes sources/openai/models-diff.json.
- Saves the sanitized listing to sources/openai/models-api-raw.json and the id list to sources/openai/model-ids.txt
  (unless --dry-run). The previous listing is kept as sources/openai/models-api-raw.prev.json.

Usage: python3 scripts/discover_openai_models.py [--dry-run]
Exit code 0 always (a diff is information, not an error); prints a one-line summary.
"""
from __future__ import annotations

import json
import sys
from datetime import datetime, timezone
from pathlib import Path

sys.path.insert(0, str(Path(__file__).resolve().parent.parent))
from scripts.live import openai_request, save_sanitized  # noqa: E402

ROOT = Path(__file__).resolve().parent.parent
RAW = ROOT / "sources/openai/models-api-raw.json"
PREV = ROOT / "sources/openai/models-api-raw.prev.json"
IDS = ROOT / "sources/openai/model-ids.txt"
DIFF = ROOT / "sources/openai/models-diff.json"


def main() -> int:
    dry = "--dry-run" in sys.argv
    previous = json.loads(RAW.read_text()) if RAW.exists() else {"data": []}
    prev_models = {m["id"]: m for m in previous.get("data", [])}

    st, body, hdrs = openai_request("GET", "/v1/models", note="discover_openai_models")
    if st != 200 or not isinstance(body, dict):
        print(f"GET /v1/models -> HTTP {st}; aborting without writing", file=sys.stderr)
        return 0
    models = sorted(body.get("data", []), key=lambda m: (m.get("created", 0), m["id"]))
    cur_models = {m["id"]: m for m in models}

    new_ids = sorted(set(cur_models) - set(prev_models))
    removed_ids = sorted(set(prev_models) - set(cur_models))
    changed = []
    for mid in sorted(set(cur_models) & set(prev_models)):
        a, b = prev_models[mid], cur_models[mid]
        for k in ("shutdown_date", "owned_by", "created"):
            if a.get(k) != b.get(k):
                changed.append({"id": mid, "field": k, "before": a.get(k), "after": b.get(k)})

    now = datetime.now(timezone.utc).strftime("%Y-%m-%dT%H:%M:%SZ")
    diff = {
        "checked_at": now,
        "previous_count": len(prev_models), "current_count": len(cur_models),
        "new": [cur_models[i] for i in new_ids],
        "removed": [prev_models[i] for i in removed_ids],
        "changed": changed,
        "dry_run": dry,
    }
    for m in diff["new"]:
        print(f"NEW      {m['id']}  (owned_by={m.get('owned_by')}, shutdown_date={m.get('shutdown_date')})")
    for m in diff["removed"]:
        print(f"REMOVED  {m['id']}  (was shutdown_date={m.get('shutdown_date')})")
    for c in changed:
        print(f"CHANGED  {c['id']}.{c['field']}: {c['before']} -> {c['after']}")
    print(f"summary: {len(cur_models)} ids live, {len(new_ids)} new, {len(removed_ids)} removed, {len(changed)} changed")

    save_sanitized(diff, DIFF)
    if not dry:
        if RAW.exists():
            PREV.write_text(RAW.read_text())
        save_sanitized({"object": body.get("object", "list"), "retrieved_at": now, "data": models}, RAW)
        IDS.write_text("\n".join(sorted(cur_models)) + "\n")
        print(f"wrote {RAW.relative_to(ROOT)}, {IDS.relative_to(ROOT)}, {DIFF.relative_to(ROOT)}")
    else:
        print(f"dry run: wrote only {DIFF.relative_to(ROOT)}")
    return 0


if __name__ == "__main__":
    sys.exit(main())