daily_papers_sync.py

#!/usr/bin/env python3
# /// script
# requires-python = ">=3.10"
# dependencies = [
#     "datasets",
#     "requests",
#     "huggingface-hub",
# ]
# ///
"""
Daily sync for huggingface/trending-papers-x dataset.
Indexes new papers and updates GitHub/project URLs via HF Papers API.

Run locally:  uv run daily_papers_sync.py
Run as HF Job: hf jobs uv run daily_papers_sync.py --secrets HF_TOKEN
"""

from __future__ import annotations

import argparse
import os
import re
import time
from dataclasses import dataclass
from datetime import datetime, timezone
from typing import Any, Optional
from urllib.parse import urlparse

import requests
from datasets import load_dataset

REPO_ID = "huggingface/trending-papers-x"
API_BASE = "https://huggingface.co/api"
DEFAULT_LIMIT = 150
DEFAULT_SLEEP_AFTER_INDEX_SECONDS = 30

# Regex patterns for arXiv ID validation
_ARXIV_URL_RE = re.compile(
    r"https?://(?:www\.)?arxiv\.org/(?:abs|pdf)/(?P<id>[^?#]+)", re.I
)
_ARXIV_NEW_RE = re.compile(r"^\d{4}\.\d{4,5}$")


@dataclass
class Candidate:
    arxiv_id: str
    github_repo: Optional[str]
    project_page: Optional[str]
    added_at: Optional[datetime]
    source: Optional[str]


@dataclass
class CandidateStats:
    link_rows_checked: int = 0
    already_indexed: int = 0
    status_errors: int = 0


def normalize_arxiv_id(value: Any) -> Optional[str]:
    """Extract and validate arXiv ID from various formats."""
    if not value:
        return None
    s = str(value).strip()

    # Extract from URL if present
    if m := _ARXIV_URL_RE.search(s):
        s = m.group("id")

    s = s.strip().rstrip("/")
    if s.lower().endswith(".pdf"):
        s = s[:-4]
    if s.lower().startswith("arxiv:"):
        s = s[6:]

    # Remove version suffix
    s = re.sub(r"v\d+$", "", s)

    # Validate new-style arXiv ID format
    if not _ARXIV_NEW_RE.fullmatch(s):
        return None

    # Validate month (positions 2-3)
    month = int(s[2:4])
    return s if 1 <= month <= 12 else None


def normalize_github_repo(value: Any) -> Optional[str]:
    """Extract and normalize GitHub repo URL."""
    if not value:
        return None
    s = str(value).strip()

    if s.startswith("git@github.com:"):
        s = f"https://github.com/{s[15:]}"
    elif s.startswith("github.com/"):
        s = f"https://{s}"

    p = urlparse(s)
    if p.scheme not in ("http", "https"):
        return None

    host = (p.netloc or "").lower().removeprefix("www.")
    if host != "github.com":
        return None

    parts = [x for x in p.path.split("/") if x]
    if len(parts) < 2:
        return None

    owner, repo = parts[0], parts[1].removesuffix(".git")
    return f"https://github.com/{owner}/{repo}"


def normalize_url(value: Any) -> Optional[str]:
    """Validate and normalize a URL."""
    if not value:
        return None
    s = str(value).strip()
    p = urlparse(s)
    return s if p.scheme in ("http", "https") and p.netloc else None


def parse_date(value: Any) -> Optional[datetime]:
    """Parse date string into datetime."""
    if isinstance(value, datetime):
        return value.replace(tzinfo=timezone.utc) if value.tzinfo is None else value
    if not value:
        return None

    for fmt in [
        "%Y-%m-%dT%H:%M:%S.%fZ",
        "%Y-%m-%dT%H:%M:%SZ",
        "%Y-%m-%dT%H:%M:%S.%f",
        "%Y-%m-%dT%H:%M:%S",
        "%Y-%m-%d",
    ]:
        try:
            dt = datetime.strptime(str(value).strip(), fmt)
            return dt.replace(tzinfo=timezone.utc)
        except ValueError:
            continue
    return None


def parse_args() -> argparse.Namespace:
    """Parse CLI args; scheduled jobs use mutating defaults."""
    parser = argparse.ArgumentParser(
        description="Index missing trending-papers-x rows with extracted links."
    )
    parser.add_argument(
        "--dry-run",
        action="store_true",
        help="Preview candidates without calling write endpoints.",
    )
    parser.add_argument(
        "--limit",
        type=int,
        default=DEFAULT_LIMIT,
        help=f"Maximum missing link-bearing rows to process. Defaults to {DEFAULT_LIMIT}.",
    )
    parser.add_argument(
        "--sleep-after-index",
        type=float,
        default=DEFAULT_SLEEP_AFTER_INDEX_SECONDS,
        help="Seconds to wait after successful indexing before updating links.",
    )
    return parser.parse_args()


def get_token() -> str:
    """Get HF token from environment or huggingface-cli."""
    if token := os.environ.get("HF_TOKEN", "").strip():
        return token
    try:
        from huggingface_hub import HfFolder

        return (HfFolder.get_token() or "").strip()
    except Exception:
        return ""


def get_paper_status(
    session: requests.Session, arxiv_id: str
) -> tuple[str, Optional[int], Optional[str]]:
    """Check whether a paper is already indexed."""
    try:
        r = session.get(f"{API_BASE}/papers/{arxiv_id}", timeout=30)
    except requests.RequestException as e:
        return "error", None, str(e)

    if r.status_code == 200:
        return "indexed", r.status_code, None
    if r.status_code == 404:
        return "missing", r.status_code, None
    return "error", r.status_code, r.text[:500]


def iter_candidates(
    session: requests.Session, limit: int
) -> tuple[list[Candidate], CandidateStats]:
    """Select live-missing trending papers that have link metadata."""
    dataset = load_dataset(REPO_ID, split="train", streaming=True)
    candidates: list[Candidate] = []
    stats = CandidateStats()
    seen: set[str] = set()

    for row in dataset:
        if len(candidates) >= limit:
            break

        arxiv_id = normalize_arxiv_id(row.get("arxiv_id") or row.get("paper_id"))
        if not arxiv_id or arxiv_id in seen:
            continue

        github_repo = normalize_github_repo(row.get("github") or row.get("github_url"))
        project_page = normalize_url(
            row.get("project_page_url") or row.get("project_page")
        )
        if not github_repo and not project_page:
            continue

        seen.add(arxiv_id)
        stats.link_rows_checked += 1
        status, http_status, error = get_paper_status(session, arxiv_id)
        if status == "indexed":
            stats.already_indexed += 1
            continue
        if status == "error":
            stats.status_errors += 1
            print(
                f"ERROR: {arxiv_id} - failed to check paper status"
                f" (status={http_status}, error={error})"
            )
            continue

        candidates.append(
            Candidate(
                arxiv_id=arxiv_id,
                github_repo=github_repo,
                project_page=project_page,
                added_at=parse_date(row.get("added_at")),
                source=row.get("source"),
            )
        )

    return candidates, stats


def index_paper(
    session: requests.Session, arxiv_id: str
) -> tuple[str, Optional[int], Optional[str]]:
    """Index a paper by arXiv ID."""
    try:
        r = session.post(
            f"{API_BASE}/papers/index", json={"arxivId": arxiv_id}, timeout=30
        )
    except requests.RequestException as e:
        return "error", None, str(e)

    if r.status_code == 200:
        return "indexed", r.status_code, None
    if r.status_code == 409:
        return "already-indexed", r.status_code, None
    return "error", r.status_code, r.text[:500]


def update_paper_links(
    session: requests.Session,
    arxiv_id: str,
    github_repo: Optional[str] = None,
    project_page: Optional[str] = None,
) -> tuple[str, Optional[int], Optional[str]]:
    """Update GitHub repo and/or project page for a paper."""
    payload = {}
    if github_repo:
        payload["githubRepo"] = github_repo
    if project_page:
        payload["projectPage"] = project_page

    if not payload:
        return "skipped", None, "no links to update"

    try:
        r = session.post(
            f"{API_BASE}/papers/{arxiv_id}/links", json=payload, timeout=30
        )
    except requests.RequestException as e:
        return "error", None, str(e)

    if r.status_code == 200:
        return "updated", r.status_code, None
    if r.status_code == 409:
        return "already-updated", r.status_code, None
    return "error", r.status_code, r.text[:500]


def main() -> None:
    args = parse_args()
    if args.limit < 1:
        print("ERROR: --limit must be greater than 0")
        exit(1)

    token = get_token()
    if not token:
        print("ERROR: HF token not found. Set HF_TOKEN or run `huggingface-cli login`.")
        exit(1)

    session = requests.Session()
    session.headers.update(
        {
            "Content-Type": "application/json",
            "Authorization": f"Bearer {token}",
        }
    )

    print(f"Dataset: {REPO_ID}")
    print(f"Limit: {args.limit} live-missing link-bearing papers")
    print(f"Mode: {'dry-run' if args.dry_run else 'apply'}")
    print("-" * 50)

    candidates, candidate_stats = iter_candidates(session, args.limit)
    if candidates:
        first_added_at = (
            candidates[0].added_at.isoformat() if candidates[0].added_at else "unknown"
        )
        last_added_at = (
            candidates[-1].added_at.isoformat()
            if candidates[-1].added_at
            else "unknown"
        )
        print(f"Candidates: {len(candidates)} ({first_added_at} -> {last_added_at})")
    else:
        print("Candidates: 0")
    print("-" * 50)

    stats = {
        "indexed": 0,
        "already_indexed": 0,
        "links": 0,
        "dry_run": 0,
        "index_errors": 0,
        "link_errors": 0,
    }

    for candidate in candidates:
        pieces = [
            candidate.arxiv_id,
            f"added_at={candidate.added_at.isoformat() if candidate.added_at else 'unknown'}",
            f"source={candidate.source or 'unknown'}",
        ]
        if candidate.github_repo:
            pieces.append(f"github={candidate.github_repo}")
        if candidate.project_page:
            pieces.append(f"project={candidate.project_page}")

        if args.dry_run:
            stats["dry_run"] += 1
            print("DRY RUN: " + " | ".join(pieces))
            continue

        index_status, index_http_status, index_error = index_paper(
            session, candidate.arxiv_id
        )
        if index_status == "indexed":
            stats["indexed"] += 1
            print(f"INDEXED: {' | '.join(pieces)}")
            if args.sleep_after_index > 0:
                time.sleep(args.sleep_after_index)
        elif index_status == "already-indexed":
            stats["already_indexed"] += 1
            print(f"ALREADY INDEXED: {' | '.join(pieces)}")
        else:
            stats["index_errors"] += 1
            print(
                f"ERROR: {candidate.arxiv_id} - failed to index"
                f" (status={index_http_status}, error={index_error})"
            )
            continue

        links_status, links_http_status, links_error = update_paper_links(
            session,
            candidate.arxiv_id,
            candidate.github_repo,
            candidate.project_page,
        )
        if links_status in {"updated", "already-updated"}:
            stats["links"] += 1
            print(f"LINKS {links_status.upper()}: {candidate.arxiv_id}")
        else:
            stats["link_errors"] += 1
            print(
                f"ERROR: {candidate.arxiv_id} - failed to update links"
                f" (status={links_http_status}, error={links_error})"
            )

    print("-" * 50)
    print(f"Link-bearing rows checked: {candidate_stats.link_rows_checked}")
    print(f"Already indexed while selecting: {candidate_stats.already_indexed}")
    print(f"Status check errors: {candidate_stats.status_errors}")
    print(f"Candidates: {len(candidates)}")
    print(f"Indexed: {stats['indexed']}")
    print(f"Already indexed: {stats['already_indexed']}")
    print(f"Links updated: {stats['links']}")
    print(f"Dry run: {stats['dry_run']}")
    print(f"Index errors: {stats['index_errors']}")
    print(f"Link errors: {stats['link_errors']}")

    if stats["index_errors"] or candidate_stats.status_errors:
        exit(1)


if __name__ == "__main__":
    main()