tg_post_min/app/bot/handlers/moderation.py

import asyncio, re
from datetime import datetime, timedelta
from telegram import Update, ChatPermissions
from telegram.constants import ChatType, ParseMode
from telegram.ext import ContextTypes

from app.db.session import get_session
from app.db.models import MessageEvent, ModerationLog
from app.moderation.engine import (
    get_policy_for_chat, compute_content_hash, redis_rate_check, redis_dupe_check,
    add_strike_and_decide_action, _domain_sets, dict_cache, _compile_dicts
)
from app.infra.metrics import MSG_PROCESSED, MSG_BLOCKED, MOD_LAT

WARN_TTL = 20

async def _autodel(ctx: ContextTypes.DEFAULT_TYPE, chat_id: int, mid: int, ttl: int = WARN_TTL):
    try:
        await asyncio.sleep(ttl)
        await ctx.bot.delete_message(chat_id=chat_id, message_id=mid)
    except Exception:
        pass

def _extract_media_ids(msg) -> list[str]:
    if msg.photo:     return [msg.photo[-1].file_id]
    if msg.video:     return [msg.video.file_id]
    if msg.animation: return [msg.animation.file_id]
    if msg.document:  return [msg.document.file_id]
    return []

async def moderate_message(update: Update, ctx: ContextTypes.DEFAULT_TYPE):
    msg = update.effective_message
    chat = update.effective_chat
    user = update.effective_user
    if chat.type not in (ChatType.GROUP, ChatType.SUPERGROUP):
        return
    if not user or user.is_bot:
        return

    MSG_PROCESSED.inc()
    with MOD_LAT.time():
        text = (getattr(msg, "text", None) or getattr(msg, "caption", None) or "")
        media_ids = _extract_media_ids(msg)
        h = compute_content_hash(text, media_ids)
        reasons = []

        with get_session() as s:
            policy = get_policy_for_chat(s, chat.id)
            if not policy:
                return

            # лог события для аналитики
            s.add(MessageEvent(chat_id=chat.id, tg_user_id=user.id, message_id=msg.message_id, content_hash=h))
            s.commit()

            # rate-limit и дубликаты через Redis (если есть)
            if policy.user_msg_per_minute > 0:
                ok = await redis_rate_check(chat.id, user.id, policy.user_msg_per_minute)
                if not ok:
                    reasons.append(f"ratelimit>{policy.user_msg_per_minute}/min")
            if policy.duplicate_window_seconds > 0:
                ok2 = await redis_dupe_check(chat.id, user.id, h, policy.duplicate_window_seconds)
                if not ok2:
                    reasons.append("duplicate")

            # дешёвые проверки
            URL_RE = re.compile(r'https?://[^\s)]+', re.IGNORECASE)
            MENTION_RE = re.compile(r'@\w+', re.UNICODE)
            links = URL_RE.findall(text or "")
            if policy.max_links >= 0 and len(links) > policy.max_links:
                reasons.append(f"links>{policy.max_links}")
            mentions = MENTION_RE.findall(text or "")
            if policy.max_mentions >= 0 and len(mentions) > policy.max_mentions:
                reasons.append(f"mentions>{policy.max_mentions}")

            wl, bl = _domain_sets(s, policy)
            from app.moderation.engine import _find_domains
            doms = _find_domains(text or "")
            if policy.use_whitelist and wl:
                bad = [d for d in doms if d not in wl]
                if bad:
                    reasons.append("not_whitelisted:" + ",".join(sorted(set(bad))))
            else:
                bad = [d for d in doms if d in bl]
                if bad:
                    reasons.append("blacklisted:" + ",".join(sorted(set(bad))))

            # словари (только если дешёвые проверки не сработали)
            if not reasons:
                ac, reg = dict_cache.get(policy.id) or _compile_dicts(s, policy)
                lo = (text or "").lower()
                matched = False
                if ac is not None:
                    for _, _ in ac.iter(lo):
                        matched = True
                        break
                if not matched:
                    for r in reg:
                        if r.search(text or ""):
                            matched = True
                            break
                if matched:
                    reasons.append("dictionary_match")

            if not reasons:
                return

            action = add_strike_and_decide_action(s, policy, chat.id, user.id)

    # Применение наказания
    MSG_BLOCKED.inc()
    performed = "none"
    try:
        if action in ("delete", "warn", "timeout", "ban"):
            try:
                await ctx.bot.delete_message(chat_id=chat.id, message_id=msg.message_id)
            except Exception:
                pass
        if action == "warn":
            warn = await msg.reply_text(
                f"⚠️ Сообщение удалено по правилам чата.\nПричины: {', '.join(reasons)}",
                parse_mode=ParseMode.HTML
            )
            ctx.application.create_task(_autodel(ctx, chat.id, warn.message_id))
            performed = "warn"
        elif action == "timeout":
            until = datetime.utcnow() + timedelta(minutes=policy.timeout_minutes)
            await ctx.bot.restrict_chat_member(chat_id=chat.id, user_id=user.id,
                                               permissions=ChatPermissions(can_send_messages=False),
                                               until_date=until)
            info = await msg.reply_text(
                f"⏳ Пользователь ограничен на {policy.timeout_minutes} минут. Причины: {', '.join(reasons)}"
            )
            ctx.application.create_task(_autodel(ctx, chat.id, info.message_id))
            performed = "timeout"
        elif action == "ban":
            await ctx.bot.ban_chat_member(chat_id=chat.id, user_id=user.id)
            info = await msg.reply_text(f"🚫 Пользователь забанен. Причины: {', '.join(reasons)}")
            ctx.application.create_task(_autodel(ctx, chat.id, info.message_id))
            performed = "ban"
        elif action == "delete":
            performed = "delete"
    finally:
        # Лог модерации
        try:
            with get_session() as s2:
                s2.add(ModerationLog(chat_id=chat.id, tg_user_id=user.id,
                                     message_id=msg.message_id,
                                     reason="; ".join(reasons), action=performed))
                s2.commit()
        except Exception:
            pass