tools/backfill_report_download_button.py

#!/usr/bin/env python3
"""Backfill the "Download submission ZIP" button into existing HTML reports.

The report generator now renders a download button in the run header (see
``cadgenbench`` ``eval/report/single_run.py``). Reports published before that
change don't have it. Rather than re-running ``evaluate`` (expensive) or even
re-rendering from artifacts, this one-off tool patches the **already-stored**
``reports/<id>.html`` files in place: it injects the button's CSS + markup at a
stable anchor in the header and re-uploads. Pure HTML string edit, no eval, no
image re-render.

Idempotent: a report that already carries the button (``class="download-zip"``)
is left untouched, so re-running is safe.

The button links the submission's ``submissions/<id>.zip`` blob URL -- the same
artifact the gallery links and the freshly-generated reports point at -- so a
patched report is visually and behaviorally identical to a fresh one.

Usage (dry-run lists what would change; nothing is written)::

    python tools/backfill_report_download_button.py

    # actually patch + re-upload (needs a write-scoped HF_TOKEN):
    python tools/backfill_report_download_button.py --apply
"""
from __future__ import annotations

import argparse
import html
import logging
import re
import sys
from pathlib import Path

# Import the Space's own read/identity helpers (this file lives in tools/, one
# level under the Space root).
sys.path.insert(0, str(Path(__file__).resolve().parent.parent))

from huggingface_hub import CommitOperationAdd, HfApi, hf_hub_download  # noqa: E402

from leaderboard import (  # noqa: E402
    HF_SUBMISSIONS_REPO,
    _load_rows_from_hub,
    _report_relative_url,
)
from submit import REPORTS_DIR, _submission_zip_url  # noqa: E402

logger = logging.getLogger(__name__)

# CSS injected before </style>. Mirrors the rules generate_html now emits so a
# patched report is byte-for-byte equivalent in the header to a fresh one.
_CSS_BLOCK = (
    "\n.run-header-top { display: flex; align-items: center; "
    "justify-content: space-between; gap: 16px; flex-wrap: wrap; }\n"
    ".run-header-top h1 { border-bottom: none; padding-bottom: 0; margin: 0; }\n"
    ".download-zip { background: #37474f; color: #fff; text-decoration: none; "
    "padding: 8px 16px; border-radius: 6px; font-size: 0.9em; "
    "font-weight: 600; white-space: nowrap; flex-shrink: 0; }\n"
    ".download-zip:hover { background: #455a64; }\n"
)

# Matches the pre-button header: the run-header div immediately followed by the
# <h1> title. Capture the title so we can re-wrap it in the flex top row.
_HEADER_RE = re.compile(
    r'(<div class="run-header">)\s*(<h1>.*?</h1>)',
    re.DOTALL,
)


def patch_html(doc: str, zip_url: str) -> str | None:
    """Return the patched HTML, or ``None`` if no change is needed/possible.

    Injects the download-button CSS before ``</style>`` and wraps the header
    title + button in a ``run-header-top`` flex row. Idempotent: returns
    ``None`` when the button is already present or the header anchor is missing.
    """
    if 'class="download-zip"' in doc:
        return None  # already patched

    href = html.escape(str(zip_url), quote=True)
    button = (
        f'<a class="download-zip" href="{href}" download rel="noopener">'
        f"&#11015; Download submission ZIP</a>"
    )

    new_doc, n = _HEADER_RE.subn(
        r'\1\n<div class="run-header-top">\n\2\n' + button + "\n</div>",
        doc,
        count=1,
    )
    if n == 0:
        return None  # header shape not recognized; skip rather than corrupt

    if "</style>" in new_doc:
        new_doc = new_doc.replace("</style>", _CSS_BLOCK + "</style>", 1)
    return new_doc


def _zip_url_for(row: dict) -> str:
    """Prefer the row's recorded blob URL; fall back to the canonical path."""
    url = row.get("submission_blob_url")
    if url:
        return str(url)
    return _submission_zip_url(row["submission_id"])


def main() -> int:
    logging.basicConfig(level=logging.INFO, format="%(message)s")
    parser = argparse.ArgumentParser(description=__doc__)
    parser.add_argument(
        "--apply", action="store_true",
        help="Re-upload patched reports (default is a dry run).",
    )
    parser.add_argument(
        "--limit", type=int, default=None,
        help="Patch at most N reports (for a cautious first pass).",
    )
    args = parser.parse_args()

    rows = _load_rows_from_hub()
    # Only completed rows from the modern pipeline have a reports/<id>.html.
    targets = [
        r for r in rows
        if r.get("submission_id")
        and _report_relative_url(
            r.get("submission_id"), r.get("status"), r.get("submission_sha256"),
        )
    ]
    logger.info("Found %d report(s) to consider.", len(targets))

    ops: list[CommitOperationAdd] = []
    skipped = 0
    for row in targets:
        sid = row["submission_id"]
        try:
            local = hf_hub_download(
                repo_id=HF_SUBMISSIONS_REPO,
                repo_type="dataset",
                filename=f"{REPORTS_DIR}/{sid}.html",
            )
        except Exception as e:  # noqa: BLE001
            logger.warning("  skip %s: could not fetch report (%s)", sid, e)
            skipped += 1
            continue
        doc = Path(local).read_text(encoding="utf-8")
        patched = patch_html(doc, _zip_url_for(row))
        if patched is None:
            logger.info("  unchanged %s (already has button or no header)", sid)
            skipped += 1
            continue
        logger.info("  patched   %s", sid)
        ops.append(
            CommitOperationAdd(
                path_in_repo=f"{REPORTS_DIR}/{sid}.html",
                path_or_fileobj=patched.encode("utf-8"),
            )
        )
        if args.limit is not None and len(ops) >= args.limit:
            break

    logger.info(
        "%d to patch, %d unchanged/skipped.", len(ops), skipped,
    )
    if not ops:
        logger.info("Nothing to do.")
        return 0
    if not args.apply:
        logger.info("Dry run -- re-run with --apply to upload.")
        return 0

    HfApi().create_commit(
        repo_id=HF_SUBMISSIONS_REPO,
        repo_type="dataset",
        operations=ops,
        commit_message="reports: backfill download-submission-zip button",
    )
    logger.info("Uploaded %d patched report(s).", len(ops))
    return 0


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