#!/usr/bin/env python3
"""doi.py — article-item pipeline: author+title → Crossref DOI → libgen availability.

For reading-list items that are JOURNAL ARTICLES (not books), the DOI is the most
reliable identifier: Crossref/OpenAlex resolve it from author+title, and libgen.vg
indexes many papers BY DOI (req=<doi> — verified 2026-09-26: Krieger
10.1111/1468-5922.12544 → exactly ed 84462696).

usage: doi.py "Lastname" "Title words..." [--year YYYY] [--no-libgen]

Steps:
1. Crossref /works?query.bibliographic=TITLE&query.author=LAST (rows=8).
2. Rank: title token overlap + year proximity (if --year).
3. For top hits with DOI: libgen req=<DOI> → count edition links (4s polite).
4. Print candidates: title | journal | author | year | DOI | LG: n (ed ids).

Limits: pre-1997 works have no DOIs (1975 Hill paper = unresolvable — fall back to
title sweeps). Conference papers in APA PsycEXTRA (10.1037/e*) are often noise.
After a hit: content-verify the downloaded file (first page: author+journal+pages
must match RAW; must be the paper, not a review of it).
"""
import json, re, sys, time, urllib.parse, urllib.request

UA = {'User-Agent': 'jung-research/1.0 (mailto:dmitry@kokorin.org)'}
LG_UA = {'User-Agent': 'Mozilla/5.0 (X11; Linux x86_64) Firefox/130.0'}


def crossref(last, title, year=None):
    q = {'query.bibliographic': title, 'query.author': last, 'rows': 8}
    if year:
        q['filter'] = f'from-pub-date:{year-2},until-pub-date:{year+2}'
    url = f"https://api.crossref.org/works?{urllib.parse.urlencode(q)}"
    req = urllib.request.Request(url, headers=UA)
    d = json.load(urllib.request.urlopen(req, timeout=40))
    out = []
    for it in d['message']['items']:
        doi = it.get('DOI')
        if not doi:
            continue
        a = it.get('author', [{}])[0]
        fam = a.get('family', '?')
        yr = (it.get('issued', {}).get('date-parts') or [['?']])[0][0]
        src = (it.get('container-title') or [''])[0]
        t = it.get('title', ['?'])[0]
        # token overlap score
        want = set(re.findall(r'\w+', title.lower()))
        got = set(re.findall(r'\w+', (t + ' ' + src).lower()))
        score = len(want & got) / max(1, len(want))
        out.append({'title': re.sub(r'<[^>]+>', '', t), 'journal': src, 'author': fam,
                    'year': yr, 'doi': doi, 'score': score})
    out.sort(key=lambda x: (-x['score'], abs((x['year'] or 0) - (year or 0)) if year else 0))
    return out


def libgen_doi(doi):
    url = f"https://libgen.vg/index.php?req={urllib.parse.quote(doi)}&res=25"
    req = urllib.request.Request(url, headers=LG_UA)
    try:
        h = urllib.request.urlopen(req, timeout=40).read().decode('utf-8', 'replace')
    except Exception as e:
        return f"ERR {e}"
    ids = sorted(set(re.findall(r'edition\.php\?id=(\d+)', h)))
    return f"{len(ids)} ({', '.join(ids[:5])})" if ids else "0"


def main():
    args = [a for a in sys.argv[1:] if not a.startswith('--')]
    flags = [a for a in sys.argv[1:] if a.startswith('--')]
    year = None
    if '--year' in sys.argv[1:]:
        year = int(sys.argv[1:][sys.argv[1:].index('--year') + 1])
    if len(args) < 2:
        print(__doc__)
        sys.exit(1)
    last, title = args[0], ' '.join(args[1:])
    hits = crossref(last, title, year)
    if not hits:
        print("(no Crossref hits with DOI)")
        return
    for h in hits[:6]:
        line = (f"[{h['score']:.2f}] {h['title'][:70]} | {h['journal'][:35]} | "
                f"{h['author']} | {h['year']} | {h['doi']}")
        if '--no-libgen' not in flags:
            time.sleep(3)
            line += f" | LG: {libgen_doi(h['doi'])}"
        print(line)


if __name__ == '__main__':
    main()