"""Download every issue labelled "feature request" in musescore/MuseScore and split it by creation date.

Writes data/explore.json (created before 2025-01-01) and data/holdout.json (created 2025-01-01 or later).
Each issue keeps: number, title, body (first 2000 characters), created_at, state, total reactions, +1 reactions,
comment count and labels. Pull requests are dropped.
"""
import json
import os
import subprocess
import sys

REPO = "musescore/MuseScore"
LABEL = "feature request"
CUTOFF = "2025-01-01T00:00:00Z"
OUT = os.environ.get("CLAIMSET_RUN_DIR") or (sys.argv[1] if len(sys.argv) > 1 else "data")


def search(created):
    """All issues matching one created-date range, via the search API (max 1000 per query)."""
    import time
    q = f'repo:{REPO} is:issue label:"{LABEL}" created:{created}'
    items, page = [], 1
    while True:
        cmd = ["gh", "api", "-X", "GET", "search/issues", "-f", f"q={q}", "-f", "per_page=100", "-f", f"page={page}"]
        res = json.loads(subprocess.run(cmd, check=True, capture_output=True, text=True).stdout)
        if res.get("incomplete_results"):
            raise RuntimeError(f"incomplete search results for {created}")
        items += res["items"]
        if res["total_count"] > 1000:
            raise RuntimeError(f"range {created} has {res['total_count']} issues; split it further")
        if len(items) >= res["total_count"] or not res["items"]:
            return items
        page += 1
        time.sleep(2.5)  # search API: 30 requests per minute


def fetch():
    ranges = ["<2019-01-01"] + [f"{y}-01-01..{y}-12-31" for y in range(2019, 2027)]
    seen, out = set(), []
    for r in ranges:
        for i in search(r):
            if i["number"] not in seen:
                seen.add(i["number"])
                out.append(i)
    return out


def slim(i):
    r = i.get("reactions") or {}
    return {"number": i["number"], "title": i["title"], "body": (i.get("body") or "")[:2000],
            "created_at": i["created_at"], "state": i["state"],
            "reactions_total": r.get("total_count", 0), "reactions_plus1": r.get("+1", 0),
            "comments": i.get("comments", 0), "labels": [l["name"] for l in i.get("labels", [])]}


def main():
    issues = [slim(i) for i in fetch() if "pull_request" not in i]
    issues.sort(key=lambda i: i["number"])
    explore = [i for i in issues if i["created_at"] < CUTOFF]
    holdout = [i for i in issues if i["created_at"] >= CUTOFF]
    os.makedirs(OUT, exist_ok=True)
    for name, part in (("explore", explore), ("holdout", holdout)):
        with open(os.path.join(OUT, f"{name}.json"), "w") as f:
            json.dump(part, f, indent=0, ensure_ascii=False)
    results = [
        {"quantity": "feature-request issues downloaded", "value": len(issues)},
        {"quantity": "feature-request issues created before 2025-01-01 (exploration set)", "value": len(explore)},
        {"quantity": "feature-request issues created on or after 2025-01-01 (held-out set)", "value": len(holdout)},
    ]
    path = os.environ.get("CLAIMSET_RESULTS")
    if path:
        with open(path, "w") as f:
            json.dump(results, f)
    print(json.dumps(results))


if __name__ == "__main__":
    main()