# ==============================================================================
# Author: Simon-Pierre Boucher <contact@spboucher.ai>
# File:   src/main.py (ka-threads)
# Desc:   Profils Meta Threads publics : la page @handle embarque le profil
#         dans des scripts JSON Relay (follower_count, biography, badge,
#         avatar) + les posts récents (caption/like_count) quand servis.
#         Extraction par regex ciblées — robuste aux réorganisations du JSON.
# ==============================================================================
from __future__ import annotations

import asyncio
import html as htmllib
import json
import re

from apify import Actor

from .net import Fetcher

PROFILE_URL = "https://www.threads.com/@{u}"

_FOLLOWERS_RE = re.compile(r'"follower_count"\s*:\s*(\d+)')
_BIO_RE = re.compile(r'"biography"\s*:\s*"((?:[^"\\]|\\.)*)"')
_NAME_RE = re.compile(r'"full_name"\s*:\s*"((?:[^"\\]|\\.)*)"')
_VERIFIED_RE = re.compile(r'"is_verified"\s*:\s*(true|false)')
_PIC_RE = re.compile(r'"profile_pic_url"\s*:\s*"((?:[^"\\]|\\.)*)"')
# liens auto-déclarés de la bio → cross-links forts côté crea-ka (§12.1)
_BIO_LINKS_RE = re.compile(r'"bio_links"\s*:\s*\[(.{0,2000}?)\]', re.S)
_LINK_URL_RE = re.compile(r'"url"\s*:\s*"(https?:(?:[^"\\]|\\.)+)"')
_PK_RE = re.compile(r'"pk"\s*:\s*"?(\d{4,})"?')
# posts : paires texte + like_count dans les payloads thread_items
_POST_RE = re.compile(
    r'"caption"\s*:\s*\{\s*"text"\s*:\s*"((?:[^"\\]|\\.)*)"[^{}]*?\}'
    r'.{0,600}?"like_count"\s*:\s*(\d+)', re.S)


async def main() -> None:
    async with Actor:
        inp = await Actor.get_input() or {}
        usernames = [u.strip().lstrip("@").lower()
                     for u in (inp.get("usernames") or []) if u and u.strip()]
        proxy = await Actor.create_proxy_configuration(
            actor_proxy_input=inp.get("proxyConfiguration"))
        fetcher = Fetcher(proxy, delay=float(inp.get("requestDelay") or 2))
        sem = asyncio.Semaphore(int(inp.get("concurrency") or 2))

        async def one(u: str) -> None:
            async with sem:
                miss = {"kind": "profile", "platform": "threads",
                        "username": u, "found": False}
                try:
                    resp = await fetcher.get(PROFILE_URL.format(u=u),
                                             session_id=u)
                except Exception as exc:
                    await Actor.push_data({**miss, "error": str(exc)[:200]})
                    return
                text = resp.text or ""
                fol = _FOLLOWERS_RE.search(text)
                if resp.status_code != 200 or not fol:
                    await Actor.push_data(
                        {**miss,
                         "error": f"shell_or_{resp.status_code}"})
                    return
                posts = []
                for m in _POST_RE.finditer(text):
                    cap = _dec(m.group(1))[:500]
                    if cap and all(p["caption"] != cap for p in posts):
                        posts.append({"caption": cap,
                                      "likes": int(m.group(2))})
                    if len(posts) >= 15:
                        break
                bio = _BIO_RE.search(text)
                name = _NAME_RE.search(text)
                ver = _VERIFIED_RE.search(text)
                pic = _PIC_RE.search(text)
                pk = _PK_RE.search(text)
                bio_links: list[str] = []
                bl = _BIO_LINKS_RE.search(text)
                if bl:
                    for m2 in _LINK_URL_RE.finditer(bl.group(1)):
                        url2 = _dec(m2.group(1))
                        if url2 not in bio_links:
                            bio_links.append(url2)
                likes = [p["likes"] for p in posts]
                top = max(posts, key=lambda p: p["likes"], default=None)
                await Actor.push_data({
                    "kind": "profile",
                    "platform": "threads",
                    "found": True,
                    "username": u,
                    "id": pk.group(1) if pk else None,
                    "full_name": _dec(name.group(1)) if name else None,
                    "biography": _dec(bio.group(1)) if bio else None,
                    "bio_links": bio_links[:5] or None,
                    "followers": int(fol.group(1)),
                    "is_verified": (ver.group(1) == "true") if ver else None,
                    "avatar": _dec(pic.group(1)) if pic else None,
                    "avg_likes": round(sum(likes) / len(likes))
                                 if likes else None,
                    "top_post": top,
                    "recent_posts": posts,
                })

        await asyncio.gather(*[one(u) for u in usernames])
        Actor.log.info(f"terminé : {len(usernames)} profils")


def _dec(raw: str) -> str:
    """Décode une chaîne échappée JSON (\\uXXXX, \\/ …)."""
    try:
        return htmllib.unescape(json.loads(f'"{raw}"'))
    except Exception:
        return raw