keyhunter-parser/cran_eternal.py

"""ETERNAL CRAN firehose: PACKAGES.gz index + source tarball scan.

Anonymous. Index = src/contrib/PACKAGES.gz (DCF, ~25k packages). Per cycle
parse index, process BATCH=120 unseen packages (alphabetical, cursor file
wraps). Each: tar.gz (cap 8MB) -> tarfile scan. seen=Package (versions
churn skipped). mark_seen ONLY on success.
"""
import os
KF_BASE = os.path.dirname(os.path.abspath(__file__))
os.makedirs(os.path.join(KF_BASE, "data"), exist_ok=True)
import gzip
import hashlib
import io
import sqlite3
import sys
import tarfile
import threading
import time

import requests

DB = os.environ.get("KF_DB") or os.path.join(KF_BASE, "data/keys.db")
LOG = os.path.join(KF_BASE, "data/cran_eternal.log")
CURSOR_FILE = os.path.join(KF_BASE, "data/cran_cursor.txt")
SEEN_FILE = os.path.join(KF_BASE, "data/cran_seen.txt")
N_THREADS = 6
BATCH = 120
TAR_CAP = 8 * 1024 * 1024
FILE_CAP = 500_000
SLEEP = 120

from key_patterns import scan_text

GARBAGE = ("xxxx", "your", "YOUR", "EXAMPLE", "example", "n0tr3al",
           "placeholder", "CHANGEME", "changeme", "dummy", "DUMMY",
           "abcdefghijklmnopqrst", "1234567890abcdef")

SCAN_EXT = (".R", ".Rmd", ".Rnw", ".py", ".js", ".json", ".yaml", ".yml",
            ".env", ".Renviron", ".csv", ".xml", ".toml", ".cfg", ".ini")

UA = {"User-Agent": "keyfarm-cran/1.0"}


def log(msg):
    line = f"[{time.strftime('%Y-%m-%d %H:%M:%S')}] {msg}\n"
    with open(LOG, "a", encoding="utf-8") as f:
        f.write(line)


conn = sqlite3.connect(DB, timeout=120, check_same_thread=False,
                       isolation_level=None)
lock = threading.Lock()
total = 0


def db_add(key, prov, repo):
    global total
    with lock:
        conn.execute(
            "INSERT INTO keys (hash, val, prov, status, found, repo) "
            "VALUES (?,?,?,?,?,?) "
            "ON CONFLICT(hash) DO UPDATE SET status='NEW', "
            "found=excluded.found, repo=excluded.repo "
            "WHERE keys.status IN ('DEAD','ERR')",
            (hashlib.sha1(key.encode()).hexdigest()[:16],
             key, prov, "NEW", "cran", repo))
        total += 1


def load_cursor():
    try:
        with open(CURSOR_FILE, encoding="utf-8") as f:
            return int(f.read().strip() or 0)
    except (FileNotFoundError, ValueError):
        return 0


def save_cursor(c):
    with open(CURSOR_FILE, "w", encoding="utf-8") as f:
        f.write(str(c))


def load_seen():
    s = set()
    try:
        with open(SEEN_FILE, encoding="utf-8") as f:
            for line in f:
                line = line.strip()
                if line:
                    s.add(line)
    except FileNotFoundError:
        pass
    return s


seen = load_seen()


def mark_seen(p):
    seen.add(p)
    with open(SEEN_FILE, "a", encoding="utf-8") as f:
        f.write(p + "\n")


def work(pkg, ver):
    """Fetch source tarball, scan. -1 on final fail."""
    n = 0
    url = (f"https://cran.r-project.org/src/contrib/"
           f"{pkg}_{ver}.tar.gz")
    for attempt in (1, 2, 3):
        try:
            r = requests.get(url, headers=UA, timeout=40)
            if r.status_code == 404:
                return 0
            if r.status_code != 200:
                raise RuntimeError(f"http {r.status_code}")
            if len(r.content) > TAR_CAP:
                return 0
            tf = tarfile.open(fileobj=io.BytesIO(r.content), mode="r:gz")
            for m in tf.getmembers():
                if not m.isfile() or m.size > FILE_CAP:
                    continue
                fn = m.name.lower()
                if not any(fn.endswith(e.lower()) for e in SCAN_EXT):
                    continue
                ef = tf.extractfile(m)
                if ef is None:
                    continue
                data = ef.read()
                if b"\x00" in data[:8192]:
                    continue  # binary archive member — no real text key has NUL
                text = data.decode("utf-8", errors="ignore")
                for prov, key in scan_text(text):
                    if isinstance(key, tuple):
                        key = key[0]
                    if any(g in key for g in GARBAGE) or len(key) < 30:
                        continue
                    db_add(key, prov, f"cran:{pkg}")
                    n += 1
            return n
        except Exception:
            if attempt == 3:
                log(f"  pkg-err {pkg}")
                return -1
            time.sleep(8)
    return -1


def cycle():
    global total
    total = 0
    # index (re-download each cycle: 1.9MB, versions bump in PACKAGES)
    for attempt in (1, 2, 3):
        try:
            r = requests.get("https://cran.r-project.org/src/contrib/"
                             "PACKAGES.gz", headers=UA, timeout=60)
            if r.status_code != 200:
                raise RuntimeError(f"http {r.status_code}")
            break
        except Exception as e:
            if attempt == 3:
                log(f"index-err {type(e).__name__}")
                return
            time.sleep(8)
    txt = gzip.decompress(r.content).decode("utf-8", "replace")
    pkgs = []
    vers = {}
    for block in txt.split("\n\n"):
        pkg = ver = None
        for line in block.splitlines():
            if line.startswith("Package: "):
                pkg = line[9:].strip()
            elif line.startswith("Version: "):
                ver = line[9:].strip()
            if pkg and ver:
                break
        if pkg and ver:
            pkgs.append(pkg)
            vers[pkg] = ver
    pkgs.sort()

    idx = load_cursor() % len(pkgs)
    items = []
    for _ in range(len(pkgs)):
        pkg = pkgs[idx]
        if pkg not in seen:
            items.append((pkg, vers[pkg]))
            if len(items) >= BATCH:
                idx = (idx + 1) % len(pkgs)
                break
        idx = (idx + 1) % len(pkgs)
    save_cursor(idx)

    ok = fail = 0
    from concurrent.futures import ThreadPoolExecutor, as_completed
    with ThreadPoolExecutor(max_workers=N_THREADS) as ex:
        futs = {ex.submit(work, p, v): p for p, v in items}
        for fut in as_completed(futs):
            p = futs[fut]
            try:
                res = fut.result()
            except Exception:
                res = -1
            if res == -1:
                fail += 1          # NOT marked seen: retried next cycle
            else:
                ok += 1
                mark_seen(p)

    log(f"cycle: total={len(pkgs)} idx={idx} items={len(items)} ok={ok} "
        f"fail={fail} +{total} keys")


def main():
    log("cran_eternal started")
    if "--once" in sys.argv:
        cycle()
        return
    while True:
        try:
            cycle()
        except Exception as e:
            log(f"cycle-err {type(e).__name__}: {e}")
        time.sleep(SLEEP)


if __name__ == "__main__":
    main()