#!/usr/bin/env python3
"""Search libgen.vg (Library Genesis). Usage: lg.sh "query" [limit]
Polite mode: one request per query."""
import sys, time, urllib.parse, urllib.request, re

UA = "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/126.0.0.0 Safari/537.36"

def fetch(url, tries=3):
    last = None
    for i in range(tries):
        try:
            req = urllib.request.Request(url, headers={"User-Agent": UA, "Accept-Language": "ru-RU,ru;q=0.9"})
            html = urllib.request.urlopen(req, timeout=90).read().decode("utf-8", "replace")
            # robot-block detection: nginx default page has no req= result table
            if 'libgen' not in html.lower()[:3000] and 'Search Result' not in html and 'result' not in html.lower()[:3000]:
                time.sleep(15)
                continue
            return html
        except Exception as e:
            last = e
            time.sleep(5 + 5 * i)
    # curl fallback (works when urllib is fingerprinted/blocked)
    import subprocess
    try:
        r = subprocess.run(["curl", "-s", "-A", UA, "--max-time", "90", url],
                           capture_output=True, text=True, timeout=120)
        if r.stdout:
            return r.stdout
    except Exception:
        pass
    raise SystemExit(f"fetch failed after {tries} tries: {last}")

def main():
    q = sys.argv[1]
    limit = int(sys.argv[2]) if len(sys.argv) > 2 else 10
    # objects[] MUST include 'f' (files): editions-only search misses files that
    # exist in libgen but are not linked to the right edition (verified 2026-09-25:
    # Solomon "Self in Transformation" found by file search, missing in edition search).
    params = {"req": q, "res": "100"}
    params["columns[]"] = ["t","a","s","y","p","i"]
    params["objects[]"] = ["f","e","s","a","p","w"]
    params["topics[]"] = ["l","c","f","a","m","r","s"]
    url = "https://libgen.vg/index.php?" + urllib.parse.urlencode(params, doseq=True)
    html = fetch(url)
    # pagination hint
    m = re.search(r'page=(\d+)', html)
    if m and int(m.group(1)) > 1:
        print(f"[more pages available, max page {m.group(1)} — rerun with &page=N]")
    # parse result rows
    rows = re.findall(r"<tr[^>]*>\s*<td[^>]*>.*?</tr>", html, re.S)
    count = 0
    for r in rows:
        tds = re.findall(r"<td[^>]*>(.*?)</td>", r, re.S)
        tds = [re.sub(r"<[^>]+>", "", t).replace("&amp;","&").strip() for t in tds]
        tds = [re.sub(r"\s+", " ", t) for t in tds]
        if len(tds) < 6:
            continue
        # first cell is usually a link with id, title is the big cell
        line = " | ".join(tds[:9])
        # file rows: surface the md5 (ads.php link) — feed to tools/lgdl.py bymd5
        md5s = re.findall(r'ads\.php\?md5=([0-9a-f]{32})', r)
        if md5s:
            line += f"  [md5={md5s[0]}]"
        print(line[:300])
        count += 1
        if count >= limit:
            break
    if count == 0:
        # maybe "No results" or error page
        m = re.search(r"No results|Ничего не найдено|Error[^<]*", html)
        print("(no results)" if m else "(unparsed — check manually)")
        print(html[:500].replace("\n", " "))

if __name__ == "__main__":
    main()