feat: Add direct navigation test and optimize _aldata polling in Anilife extractor.

lib/camoufox_anilife.py

@@ -13,40 +13,34 @@ import re
 import os
 
 async def _wait_for_aldata(page, timeout=8):
-    """_aldata 변수가 나타날 때까지 폴링 (최대 timeout초)"""
+    """_aldata 변수가 나타날 때까지 고속 폴링 (50ms)"""
     start_time = asyncio.get_event_loop().time()
     while asyncio.get_event_loop().time() - start_time < timeout:
         try:
-            # 1. JS 변수 확인
+            # 1. JS 변수 확인 (가장 빠름)
             aldata = await page.evaluate("typeof _aldata !== 'undefined' ? _aldata : null")
             if aldata:
                 return aldata, "JS"
             
-            # 2. HTML 소스 패턴 확인
+            # 2. HTML 소스 패턴 확인 (커밋 직후에 바로 걸릴 수 있음)
             html = await page.content()
             match = re.search(r'_aldata\s*=\s*["\']([A-Za-z0-9+/=]+)["\']', html)
             if match:
                 return match.group(1), "HTML"
         except:
             pass
-        await asyncio.sleep(0.2)
+        await asyncio.sleep(0.05) # 50ms로 단축
     return None, None
 
-async def _run_browser(browser, detail_url, episode_num, result):
+async def _run_browser(browser, detail_url, episode_num, result, provider_url=None):
-    """최적화된 브라우저 작업 수행"""
+    """세션 유지 + 직접 리다이렉트 방식 (클릭/새창 없음)"""
     start_time_all = asyncio.get_event_loop().time()
     page = await browser.new_page()
     
-    # 공격적 리소스 및 트래킹 차단
+    # 리소스 차단 (스크립트는 허용)
     async def intercept(route):
-        req_url = route.request.url.lower()
         resource_type = route.request.resource_type
-        
+        if resource_type in ["image", "media", "font", "stylesheet"]:
-        # 차단 목록: 이미지, 미디어, 폰트, 스타일시트, 분석/광고 스크립트
-        block_patterns = ["google-analytics", "googletagmanager", "facebook.net", "ads"]
-        block_types = ["image", "media", "font", "stylesheet"]
-        
-        if resource_type in block_types or any(p in req_url for p in block_patterns):
             await route.abort()
         else:
             await route.continue_()
@@ -54,78 +48,55 @@ async def _run_browser(browser, detail_url, episode_num, result):
     await page.route("**/*", intercept)
     
     try:
-        # 1. Detail 페이지 이동
+        # 1. Detail 페이지 이동 → 세션/쿠키 획득
         t_nav_start = asyncio.get_event_loop().time()
-        print(f"1. Navigating: {detail_url}", file=sys.stderr)
+        print(f"1. Session: {detail_url}", file=sys.stderr)
-        await page.goto(detail_url, wait_until="commit", timeout=15000)
+        await page.goto(detail_url, wait_until="commit", timeout=10000)
-        print(f"   Navigation took: {round(asyncio.get_event_loop().time() - t_nav_start, 2)}s", file=sys.stderr)
+        print(f"   Done in {round(asyncio.get_event_loop().time() - t_nav_start, 2)}s", file=sys.stderr)
         
-        # 2. 에피소드 링크 찾기 및 클릭
+        # 2. 에피소드 링크에서 href 추출 (클릭 X)
         t_find_start = asyncio.get_event_loop().time()
-        print(f"2. Searching episode {episode_num}...", file=sys.stderr)
+        print(f"2. Finding ep {episode_num} link...", file=sys.stderr)
-        episode_link = None
+        
-        for _ in range(20): # 약 4초
-            try:
-                # epl-num 텍스트 매칭
         episode_link = page.locator(f'a:has(.epl-num:text("{episode_num}"))').first
-                if await episode_link.is_visible():
+        for _ in range(20):
-                    break
+            if await episode_link.is_visible(): break
+            await asyncio.sleep(0.1)
         
-                # 대체: provider 링크
+        # 클릭 방식으로 네비게이션 (직접 URL 접근은 사이트에서 막힘)
-                links = await page.locator('a[href*="/ani/provider/"]').all()
+        print(f"   Link found in {round(asyncio.get_event_loop().time() - t_find_start, 2)}s. Clicking...", file=sys.stderr)
-                for link in links:
-                    if episode_num in await link.inner_text():
-                        episode_link = link
-                        break
-                if episode_link: break
-            except: pass
-            await asyncio.sleep(0.2)
-        
-        if not episode_link:
-            result["error"] = f"Episode {episode_num} not found"
-            return result
-
-        print(f"   Finding link took: {round(asyncio.get_event_loop().time() - t_find_start, 2)}s", file=sys.stderr)
-
-        # 3. 에피소드 클릭
-        t_click_start = asyncio.get_event_loop().time()
         await episode_link.click()
         
-        # 4. _aldata 추출 (최대 6초 폴링)
+        # 3. _aldata 추출 (고속 폴링)
+        print("3. Extracting _aldata...", file=sys.stderr)
         aldata, source = await _wait_for_aldata(page, timeout=6)
         
-        if aldata:
+        # 버튼 클릭 폴백
-            elapsed = asyncio.get_event_loop().time() - start_time_all
+        if not aldata:
-            result.update({
+            print("   Trying player button...", file=sys.stderr)
-                "aldata": aldata, "success": True, 
-                "elapsed": round(elapsed, 2), "source": source
-            })
-            print(f"   SUCCESS! Extracted via {source} in {result['elapsed']}s", file=sys.stderr)
-            return result
-            
-        # 5. 최후의 수단: 플레이어 버튼 클릭 시도
-        print(f"   Initial extraction failed ({round(asyncio.get_event_loop().time() - t_click_start, 2)}s). Trying player button...", file=sys.stderr)
             btn = page.locator('a[onclick*="moveCloudvideo"], a[onclick*="moveJawcloud"]').first
-        if await btn.is_visible(timeout=1500):
+            for _ in range(20):  # 2초 대기
-            await btn.click()
+                if await btn.is_visible(): break
+                await asyncio.sleep(0.1)
+            if await btn.is_visible():
+                await btn.click(force=True)
                 aldata, source = await _wait_for_aldata(page, timeout=4)
+                if aldata: source = f"{source}-btn"
+        
         if aldata:
             elapsed = asyncio.get_event_loop().time() - start_time_all
-                result.update({
+            result.update({"aldata": aldata, "success": True, "elapsed": round(elapsed, 2), "source": source})
-                    "aldata": aldata, "success": True, 
+            print(f"   SUCCESS in {result['elapsed']}s ({source})", file=sys.stderr)
-                    "elapsed": round(elapsed, 2), "source": f"{source}-player"
-                })
-                print(f"   SUCCESS! Got aldata via player in {result['elapsed']}s", file=sys.stderr)
             return result
         
-        result["error"] = "Aldata extraction failed"
+        result["error"] = "Failed to extract aldata"
             
     finally:
         await page.close()
     
     return result
 
-async def extract_aldata(detail_url: str, episode_num: str) -> dict:
+async def extract_aldata(detail_url: str, episode_num: str, provider_url: str = None) -> dict:
     """AsyncCamoufox Stealth-Headless mode"""
     try:
         from camoufox.async_api import AsyncCamoufox
@@ -137,7 +108,7 @@ async def extract_aldata(detail_url: str, episode_num: str) -> dict:
     try:
         # Camoufox는 headless=True에서도 강력한 스텔스를 제공함 (Xvfb 오버헤드 불필요)
         async with AsyncCamoufox(headless=True) as browser:
-            return await _run_browser(browser, detail_url, episode_num, result)
+            return await _run_browser(browser, detail_url, episode_num, result, provider_url)
             
     except Exception as e:
         result["error"] = str(e)
@@ -148,9 +119,13 @@ if __name__ == "__main__":
     if len(sys.argv) < 3:
         sys.exit(1)
     
+    detail_url = sys.argv[1]
+    episode_num = sys.argv[2]
+    provider_url = sys.argv[3] if len(sys.argv) > 3 else None
+    
     # stdout에는 오직 JSON만 출력하도록 보장
     try:
-        res = asyncio.run(extract_aldata(sys.argv[1], sys.argv[2]))
+        res = asyncio.run(extract_aldata(detail_url, episode_num, provider_url))
         # 최종 JSON 결과 출력
         print(json.dumps(res, ensure_ascii=False))
     except Exception as e:

lib/test_camoufox_direct.py

@@ -0,0 +1,48 @@
+import asyncio
+import sys
+import json
+import re
+from camoufox.async_api import AsyncCamoufox
+
+async def test_extraction(detail_url, provider_url):
+    print(f"Testing direct navigation with Referer...")
+    async with AsyncCamoufox(headless=True) as browser:
+        page = await browser.new_page()
+        
+        # 1. Detail page (establish Referer)
+        print(f"1. Estabilishing Referer: {detail_url}")
+        t1 = asyncio.get_event_loop().time()
+        await page.goto(detail_url, wait_until="commit")
+        print(f"   Took {round(asyncio.get_event_loop().time() - t1, 2)}s")
+        
+        # 2. Same-session direct navigation to provider
+        print(f"2. Navigating directly to provider: {provider_url}")
+        t2 = asyncio.get_event_loop().time()
+        await page.goto(provider_url, wait_until="commit")
+        print(f"   Took {round(asyncio.get_event_loop().time() - t2, 2)}s")
+        
+        # 3. Check for aldata
+        html = await page.content()
+        final_url = page.url
+        print(f"Final URL: {final_url}")
+        
+        if "google.com" in final_url:
+            print("FAILED: Redirected to Google (Bot detection triggered)")
+        else:
+            match = re.search(r'_aldata\s*=\s*["\']([A-Za-z0-9+/=]+)["\']', html)
+            if match:
+                print("SUCCESS: Got aldata via direct navigation!")
+            else:
+                print("FAILED: Aldata not found in HTML")
+
+if __name__ == "__main__":
+    # Sample URLs for testing
+    # Note: These are placeholders, I will use real ones if available from logs
+    d_url = "https://anilife.live/detail/id/2967"
+    p_url = "https://anilife.live/ani/provider/31db6215-62bb-420a-8d18-9717013854eb"
+    
+    if len(sys.argv) > 2:
+        d_url = sys.argv[1]
+        p_url = sys.argv[2]
+        
+    asyncio.run(test_extraction(d_url, p_url))

mod_anilife.py

@@ -1281,7 +1281,7 @@ class AniLifeQueueEntity(FfmpegQueueEntity):
                 import json as json_module
                 
                 # 셋업 확인 (이미 완료되었으면 즉시 반환, 아니면 대기)
-                if not self.ensure_camoufox_installed():
+                if not self.module_logic.ensure_camoufox_installed():
                     logger.error("Camoufox installation failed. Cannot proceed.")
                     return
                 
@@ -1291,28 +1291,51 @@ class AniLifeQueueEntity(FfmpegQueueEntity):
                 # detail_url과 episode_num 추출
                 detail_url = self.info.get("ep_url", f"https://anilife.live/detail/id/{self.info.get('content_code', '')}")
                 episode_num = str(self.info.get("ep_num", "1"))
+                provider_url = self.info.get("va") # 직접 진입용 프로바이더 URL
+                if provider_url and provider_url.startswith("/"):
+                    provider_url = f"https://anilife.live{provider_url}"
                 
                 logger.debug(f"Running Camoufox subprocess: {script_path}")
-                logger.debug(f"Detail URL: {detail_url}, Episode: {episode_num}")
+                logger.debug(f"Detail URL: {detail_url}, Episode: {episode_num}, Provider: {provider_url}")
                 
-                # subprocess로 Camoufox 스크립트 실행
+                # subprocess로 Camoufox 스크립트 실행 (stderr 실시간 로그 연동)
-                result = subprocess.run(
+                cmd = [sys.executable, script_path, detail_url, episode_num]
-                    [sys.executable, script_path, detail_url, episode_num],
+                if provider_url:
-                    capture_output=True,
+                    cmd.append(provider_url)
-                    text=True,
+                
-                    timeout=120  # 120초 타임아웃
+                process = subprocess.Popen(
+                    cmd,
+                    stdout=subprocess.PIPE,
+                    stderr=subprocess.PIPE,
+                    text=True
                 )
                 
-                if result.returncode != 0:
+                # stderr를 실시간으로 logger.info에 기록 (진단 가시성 확보)
-                    logger.error(f"Camoufox subprocess failed: {result.stderr}")
+                stdout_data = []
-                    raise Exception(f"Subprocess error: {result.stderr}")
+                import threading
+                def log_stderr(pipe):
+                    for line in iter(pipe.readline, ''):
+                        if line.strip():
+                            logger.info(f"[Camoufox] {line.strip()}")
                 
-                # JSON 결과 파싱 (엄격한 분리를 통해 stdout에는 JSON만 남음)
+                stderr_thread = threading.Thread(target=log_stderr, args=(process.stderr,))
+                stderr_thread.start()
+                
+                # stdout 캡처 (JSON 결과)
+                for line in iter(process.stdout.readline, ''):
+                    stdout_data.append(line)
+                
+                process.wait(timeout=120)
+                stderr_thread.join(timeout=5)
+                
+                stdout_full = "".join(stdout_data)
+                
+                # JSON 결과 파싱
                 try:
-                    cf_result = json_module.loads(result.stdout)
+                    cf_result = json_module.loads(stdout_full)
                 except json_module.JSONDecodeError as e:
                     logger.error(f"Failed to parse Camoufox result: {e}")
-                    logger.error(f"Raw stdout: {result.stdout}")
+                    logger.debug(f"Raw stdout: {stdout_full}")
                     return
                 
                 elapsed = cf_result.get("elapsed", "?")