真实链接测试

douyin_parser/main.py

@@ -55,6 +55,7 @@ class DouyinParser:
         max_retries = 3  # 最大重试次数
         retry_delay = 2  # 重试延迟秒数
         proxies = {"http": self.http_proxy, "https": self.http_proxy} if self.http_proxy else None
+        redirect_history = []
 
         for retry in range(max_retries):
             try:
@@ -66,30 +67,42 @@ class DouyinParser:
                     'Range': 'bytes=0-'
                 }
 
-                response = requests.get(video_url, headers=headers, proxies=proxies, allow_redirects=True, timeout=60)
+                current_url = video_url
-                if response.status_code in [200, 206]:
+                while True:
-                    history = [resp.url for resp in response.history]
+                    response = requests.get(current_url, headers=headers, proxies=proxies, allow_redirects=False,
-                    real_url = response.url
+                                            timeout=60)
 
-                    if history:
+                    if response.status_code in [301, 302, 303, 307, 308]:  # 处理重定向
-                        self.LOG.info("[抖音] 重定向历史: %s", history)
+                        new_url = response.headers.get('Location')
+                        if not new_url:
+                            self.LOG.info("[抖音] 发现重定向但没有 Location 头部，停止解析")
+                            break
+                        if not new_url.startswith("http"):
+                            from urllib.parse import urljoin
+                            new_url = urljoin(current_url, new_url)  # 处理相对路径重定向
 
-                    if real_url != video_url and ('v3-' in real_url.lower() or 'douyinvod.com' in real_url.lower()):
+                        redirect_history.append(new_url)
-                        self.LOG.info("[抖音] 成功获取真实链接: %s", real_url)
+                        self.LOG.info("[抖音] 发现重定向: %s -> %s", current_url, new_url)
-                        return real_url
+                        current_url = new_url
                     else:
-                        self.LOG.warning("[抖音] 未能获取到真实视频链接，准备重试")
+                        break  # 没有进一步重定向，停止循环
-                        if retry < max_retries - 1:
+
-                            time.sleep(retry_delay)
+                real_url = current_url
-                            continue
+                self.LOG.info("[抖音] 最终获取到的链接: %s", real_url)
-                        return video_url
+
+                if redirect_history:
+                    self.LOG.info("[抖音] 重定向历史: %s", redirect_history)
+
+                if real_url != video_url and ('v3-' in real_url.lower() or 'douyinvod.com' in real_url.lower()):
+                    self.LOG.info("[抖音] 成功获取真实链接: %s", real_url)
+                    return real_url
                 else:
-                    self.LOG.error("[抖音] 获取视频真实链接失败, 状态码: %d", response.status_code)
+                    self.LOG.info("[抖音] 未能获取到符合预期的视频链接，准备重试")
-                    self.LOG.debug("[抖音] 响应头: %s", response.headers)
                     if retry < max_retries - 1:
                         time.sleep(retry_delay)
                         continue
                     return video_url
+
             except Exception as e:
                 self.LOG.error("[抖音] 获取真实链接失败: %s (第%d次尝试)", str(e), retry + 1)
                 if retry < max_retries - 1: