v0.3.4: update PpomppuScraper.kt

This commit is contained in:
sanjeok77 committed 2026-08-25 01:58:42 +00:00
1 parent 0597a002cd
commit cf2d7c615a
1 file changed
+12 -12
@@ -3,6 +3,7 @@ package com.hotdeal.alarm.data.remote.scraper
import android.util.Log import android.util.Log
import com.hotdeal.alarm.domain.model.HotDeal import com.hotdeal.alarm.domain.model.HotDeal
import kotlinx.coroutines.Dispatchers import kotlinx.coroutines.Dispatchers
import kotlinx.coroutines.async
import kotlinx.coroutines.delay import kotlinx.coroutines.delay
import kotlinx.coroutines.withContext import kotlinx.coroutines.withContext
import okhttp3.OkHttpClient import okhttp3.OkHttpClient
@@ -13,7 +14,7 @@ import kotlin.random.Random
/** /**
* 뽐뿌 스크래퍼 * 뽐뿌 스크래퍼
* - hotlist_flag=999 페이지에서 인기글 ID를 수집하여 HOT 뱃지 표시 * - hotlist_flag=999 페이지와 일반 게시판을 병렬 스크래핑하여 HOT 뱃지 신속 감지
* - 기존 hotpop_bg_color CSS 클래스도 폴백으로 감지 * - 기존 hotpop_bg_color CSS 클래스도 폴백으로 감지
*/ */
class PpomppuScraper(client: OkHttpClient) : BaseScraper(client) { class PpomppuScraper(client: OkHttpClient) : BaseScraper(client) {
@@ -27,15 +28,13 @@ class PpomppuScraper(client: OkHttpClient) : BaseScraper(client) {
override suspend fun scrape(board: String): Result<List<HotDeal>> = withContext(Dispatchers.IO) { override suspend fun scrape(board: String): Result<List<HotDeal>> = withContext(Dispatchers.IO) {
try { try {
// 1단계: 인기글(hotlist) 페이지에서 인기 게시물 ID 목록 수집
val hotPostIds = fetchHotPostIds(board)
Log.d("Ppomppu", "인기글 ID ${hotPostIds.size}개 수집됨: $hotPostIds")
// 2단계: 일반 게시판 페이지 스크래핑
val url = getBoardUrl(board) val url = getBoardUrl(board)
Log.d("Ppomppu", "스크래핑 시작: $url") Log.d("Ppomppu", "스크래핑 시작: $url")
val delayTime = Random.nextLong(2000, 4000) // 1. 인기글 목록과 일반 게시판을 병렬(async)로 동시에 요청하여 딜레이 최소화
val hotPostIdsDeferred = async { fetchHotPostIds(board) }
val delayTime = Random.nextLong(500, 1500)
delay(delayTime) delay(delayTime)
val userAgent = getRandomUserAgent() val userAgent = getRandomUserAgent()
@@ -45,17 +44,20 @@ class PpomppuScraper(client: OkHttpClient) : BaseScraper(client) {
.header("Accept-Language", "ko-KR,ko;q=0.9,en-US;q=0.8,en;q=0.7") .header("Accept-Language", "ko-KR,ko;q=0.9,en-US;q=0.8,en;q=0.7")
.header("Accept-Charset", "utf-8,ISO-8859-1") .header("Accept-Charset", "utf-8,ISO-8859-1")
.header("Referer", "https://www.ppomppu.co.kr/") .header("Referer", "https://www.ppomppu.co.kr/")
.timeout(30000) .timeout(12000)
.followRedirects(true) .followRedirects(true)
.get() .get()
val hotPostIds = hotPostIdsDeferred.await()
Log.d("Ppomppu", "인기글 ID ${hotPostIds.size}개 수집됨: $hotPostIds")
val deals = mutableListOf<HotDeal>() val deals = mutableListOf<HotDeal>()
val rowElements: Elements = doc.select("tr.baseList") val rowElements: Elements = doc.select("tr.baseList")
Log.d("Ppomppu", "찾은 행 요소: ${rowElements.size}개") Log.d("Ppomppu", "찾은 행 요소: ${rowElements.size}개")
var count = 0 var count = 0
rowElements.forEach { row -> rowElements.forEach { row ->
if (count >= 20) return@forEach if (count >= 25) return@forEach
try { try {
val titleElement = row.selectFirst("a.baseList-title") ?: return@forEach val titleElement = row.selectFirst("a.baseList-title") ?: return@forEach
@@ -108,14 +110,12 @@ class PpomppuScraper(client: OkHttpClient) : BaseScraper(client) {
val hotUrl = "${baseUrl}zboard.php?id=$board&hotlist_flag=999" val hotUrl = "${baseUrl}zboard.php?id=$board&hotlist_flag=999"
Log.d("Ppomppu", "인기글 페이지 조회: $hotUrl") Log.d("Ppomppu", "인기글 페이지 조회: $hotUrl")
delay(Random.nextLong(1000, 2000))
val doc = Jsoup.connect(hotUrl) val doc = Jsoup.connect(hotUrl)
.userAgent(getRandomUserAgent()) .userAgent(getRandomUserAgent())
.header("Accept", "text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8") .header("Accept", "text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8")
.header("Accept-Language", "ko-KR,ko;q=0.9,en-US;q=0.8,en;q=0.7") .header("Accept-Language", "ko-KR,ko;q=0.9,en-US;q=0.8,en;q=0.7")
.header("Referer", "https://www.ppomppu.co.kr/") .header("Referer", "https://www.ppomppu.co.kr/")
.timeout(15000) .timeout(8000)
.followRedirects(true) .followRedirects(true)
.get() .get()