Repository Public
1# ==============================================================================2# Author: Simon-Pierre Boucher <contact@spboucher.ai>3# File: src/main.py (ka-facebook)4# Desc: Pages Facebook publiques — extraction best-effort du HTML servi aux5# navigateurs non connectés : compteur d'abonnés (JSON embarqué6# follower_count / texte « N followers/abonnés »), nom (og:title),7# avatar (og:image), catégorie. EXPÉRIMENTAL : FB mure agressivement ;8# un résultat found=False n'est pas une erreur du pipeline.9# ==============================================================================10from __future__ import annotations1112import asyncio13import html as htmllib14import re1516from apify import Actor1718from .net import Fetcher1920PAGE_URL = "https://www.facebook.com/{u}"2122_OG_RE = {23 "title": re.compile(r'<meta property="og:title" content="([^"]*)"'),24 "image": re.compile(r'<meta property="og:image" content="([^"]*)"'),25 "description": re.compile(26 r'<meta (?:property="og:description"|name="description") '27 r'content="([^"]*)"'),28}29_FOLLOWER_RES = (30 re.compile(r'"follower_count"\s*:\s*(\d+)'),31 re.compile(r'"global_likers_count"\s*:\s*(\d+)'),32)33# texte « 1,2 M followers » / « 12 k abonnés » / « 4 016 971 mentions J’aime »34_TEXT_COUNT_RE = re.compile(35 r'([\d][\d\s .,]*)\s*([KkMm]?)\s*'36 r'(?:followers|abonnés|mentions\s+J’aime|J’aime|likes)',37 re.I)38_CATEGORY_RE = re.compile(r'"category_name"\s*:\s*"([^"]+)"')39_VERIFIED_RE = re.compile(r'"is_verified"\s*:\s*(true|false)')404142def parse_text_count(text: str) -> int | None:43 m = _TEXT_COUNT_RE.search(text or "")44 if not m:45 return None46 num = m.group(1).replace(" ", "").replace(" ", "").replace(",", ".")47 try:48 val = float(num)49 except ValueError:50 return None51 unit = (m.group(2) or "").lower()52 if unit == "k":53 val *= 1_00054 elif unit == "m":55 val *= 1_000_00056 return int(val)575859async def main() -> None:60 async with Actor:61 inp = await Actor.get_input() or {}62 usernames = []63 for raw in (inp.get("usernames") or []):64 raw = (raw or "").strip()65 if not raw:66 continue67 raw = re.sub(r"^https?://(www\.|m\.)?facebook\.com/", "", raw)68 usernames.append(raw.strip("/").split("?")[0])69 proxy = await Actor.create_proxy_configuration(70 actor_proxy_input=inp.get("proxyConfiguration"))71 fetcher = Fetcher(proxy, delay=float(inp.get("requestDelay") or 2))72 sem = asyncio.Semaphore(int(inp.get("concurrency") or 2))7374 async def one(u: str) -> None:75 async with sem:76 miss = {"kind": "profile", "platform": "facebook",77 "username": u, "found": False}78 try:79 resp = await fetcher.get(PAGE_URL.format(u=u),80 session_id=u)81 except Exception as exc:82 await Actor.push_data({**miss, "error": str(exc)[:200]})83 return84 text = resp.text or ""85 if resp.status_code != 200 or len(text) < 5000:86 await Actor.push_data(87 {**miss, "error": f"http_{resp.status_code}"})88 return89 followers = None90 for rx in _FOLLOWER_RES:91 m = rx.search(text)92 if m:93 followers = int(m.group(1))94 break95 og = {k: htmllib.unescape(m.group(1)) if (m := rx.search(text))96 else None for k, rx in _OG_RE.items()}97 if followers is None:98 followers = parse_text_count(og.get("description") or "")99 if followers is None: # texte « N followers/abonnés » du corps100 followers = parse_text_count(text)101 if followers is None and not og.get("title"):102 await Actor.push_data({**miss, "error": "login_wall"})103 return104 cat = _CATEGORY_RE.search(text)105 ver = _VERIFIED_RE.search(text)106 await Actor.push_data({107 "kind": "profile",108 "platform": "facebook",109 "found": True,110 "username": u,111 "full_name": og.get("title"),112 "biography": og.get("description"),113 "followers": followers,114 "category": cat.group(1) if cat else None,115 "is_verified": (ver.group(1) == "true") if ver else None,116 "avatar": og.get("image"),117 "og_description": og.get("description"),118 })119120 await asyncio.gather(*[one(u) for u in usernames])121 Actor.log.info(f"terminé : {len(usernames)} pages")122