infohliadka/pages/api/sources/check.ts

import type { NextApiRequest, NextApiResponse } from 'next'
import { db, schema } from '../../../lib/db/connection'
import { eq, and, sql } from 'drizzle-orm'
import { rateLimit, getRateLimitHeaders } from '../../../lib/rate-limiter'
import { cache, getCacheKey } from '../../../lib/cache'

type CheckResponse = {
  is_problematic: boolean
  risk_level: number
  categories: string[]
  message: string
  source_count: number
}

function extractDomain(url: string): string {
  try {
    // Clean up the URL first
    let cleanUrl = url.trim()
    if (!cleanUrl.startsWith('http://') && !cleanUrl.startsWith('https://')) {
      cleanUrl = 'https://' + cleanUrl
    }
    
    const urlObj = new URL(cleanUrl)
    let domain = urlObj.hostname.toLowerCase()
    
    // Remove common prefixes
    domain = domain.replace(/^www\./, '')
    domain = domain.replace(/^m\./, '')
    domain = domain.replace(/^mobile\./, '')
    domain = domain.replace(/^amp\./, '')
    
    // Handle subdomains for known patterns - extract main domain for common TLDs
    if (domain.includes('.')) {
      const parts = domain.split('.')
      if (parts.length > 2) {
        const tld = parts[parts.length - 1]
        const sld = parts[parts.length - 2]
        // Extract main domain for common TLDs
        if (['com', 'org', 'net', 'sk', 'cz', 'hu', 'pl', 'eu', 'info'].includes(tld)) {
          domain = `${sld}.${tld}`
        }
      }
    }
    
    return domain
  } catch {
    return ''
  }
}

export default async function handler(
  req: NextApiRequest,
  res: NextApiResponse<CheckResponse | { error: string }>
) {
  if (req.method !== 'GET') {
    return res.status(405).json({ error: 'Method not allowed' })
  }

  // Rate limiting
  const clientIp = req.headers['x-forwarded-for'] || req.connection?.remoteAddress || 'unknown'
  const rateLimitResult = rateLimit(clientIp.toString())
  
  const headers = getRateLimitHeaders(rateLimitResult)
  Object.entries(headers).forEach(([key, value]) => {
    res.setHeader(key, value)
  })
  
  if (!rateLimitResult.allowed) {
    return res.status(429).json({ error: 'Too many requests' })
  }

  const { url } = req.query

  if (!url || typeof url !== 'string') {
    return res.status(400).json({ error: 'URL parameter is required' })
  }

  const domain = extractDomain(url)
  if (!domain) {
    return res.status(400).json({ error: 'Invalid URL format' })
  }

  // Check cache first
  const cacheKey = getCacheKey('domain_check', domain)
  const cachedResult = cache.get<CheckResponse>(cacheKey)
  
  if (cachedResult) {
    return res.status(200).json(cachedResult)
  }

  try {
    const sources = await db
      .select({
        id: schema.sources.id,
        riskLevel: schema.sources.riskLevel,
        categories: sql<string>`string_agg(${schema.categories.name}, ',')`
      })
      .from(schema.sources)
      .leftJoin(schema.sourceCategories, eq(schema.sources.id, schema.sourceCategories.sourceId))
      .leftJoin(schema.categories, eq(schema.sourceCategories.categoryId, schema.categories.id))
      .where(
        and(
          eq(schema.sources.domain, domain),
          eq(schema.sources.status, 'verified')
        )
      )
      .groupBy(schema.sources.id, schema.sources.riskLevel)

    let result: CheckResponse

    if (sources.length === 0) {
      result = {
        is_problematic: false,
        risk_level: 0,
        categories: [],
        message: 'Stránka nie je v našej databáze problematických zdrojov',
        source_count: 0
      }
    } else {
      const maxRiskLevel = Math.max(...sources.map(s => s.riskLevel || 0))
      const allCategories = sources
        .map(s => s.categories)
        .filter(Boolean)
        .join(',')
        .split(',')
        .filter(Boolean)

      const uniqueCategories = Array.from(new Set(allCategories))

      let message = ''
      if (maxRiskLevel >= 4) {
        message = 'VYSOKÉ RIZIKO: Táto stránka šíri nebezpečné obsahy'
      } else if (maxRiskLevel >= 3) {
        message = 'STREDNÉ RIZIKO: Táto stránka môže obsahovať problematické informácie'
      } else {
        message = 'NÍZKE RIZIKO: Táto stránka je označená ako problematická'
      }

      result = {
        is_problematic: true,
        risk_level: maxRiskLevel,
        categories: uniqueCategories,
        message,
        source_count: sources.length
      }
    }

    // Cache the result (5 minutes for non-problematic, 15 minutes for problematic)
    const cacheTtl = result.is_problematic ? 900 : 300
    cache.set(cacheKey, result, cacheTtl)

    return res.status(200).json(result)

  } catch (error) {
    console.error('Database error:', error)
    return res.status(500).json({ error: 'Internal server error' })
  }
}