feat: 快手关键词搜索存储CSV完成

media_platform/kuaishou/client.py

@@ -31,13 +31,6 @@ class KuaiShouClient:
         self.cookie_dict = cookie_dict
         self.graphql = KuaiShouGraphQL()
 
-    async def _pre_headers(self, url: str, data=None):
-        self.headers = {
-            "cookie":"kpf=PC_WEB; clientid=3; did=web_6e79b79fdeac627f5cf52f08cab4e6bd; kpn=KUAISHOU_VISION",
-            "content-type":"application/json"
-        }
-        return self.headers
-
     async def request(self, method, url, **kwargs) -> Dict:
         async with httpx.AsyncClient(proxies=self.proxies) as client:
             response = await client.request(
@@ -55,8 +48,7 @@ class KuaiShouClient:
         if isinstance(params, dict):
             final_uri = (f"{uri}?"
                          f"{urlencode(params)}")
-        headers = await self._pre_headers(final_uri)
-        return await self.request(method="GET", url=f"{self._host}{final_uri}", headers=headers)
+        return await self.request(method="GET", url=f"{self._host}{final_uri}", headers=self.headers)
 
     async def post(self, uri: str, data: dict) -> Dict:
         json_str = json.dumps(data, separators=(',', ':'), ensure_ascii=False)
@@ -81,12 +73,12 @@ class KuaiShouClient:
 
     async def search_info_by_keyword(self, keyword: str, pcursor: str):
         """
-        KuaiShou Web Search API
+        KuaiShou web search api
         :param keyword: search keyword
         :param pcursor: limite page curson
         :return:
         """
-        params = {
+        post_data = {
             "operationName": "visionSearchPhoto",
             "variables": {
                 "keyword": keyword,
@@ -95,8 +87,21 @@ class KuaiShouClient:
             },
             "query": self.graphql.get("search_query")
         }
-        return await self.post("", params)
+        return await self.post("", post_data)
 
 
-    async def get_video_info(self, video_id: str) -> Dict:
-        pass
+    async def get_video_info(self, photo_id: str) -> Dict:
+        """
+        Kuaishou web video detail api
+        :param photo_id:
+        :return:
+        """
+        post_data = {
+            "operationName": "visionVideoDetail",
+            "variables": {
+                "photoId": photo_id,
+                "page": "search"
+            },
+            "query": self.graphql.get("video_detail")
+        }
+        return await self.post("", post_data)

media_platform/kuaishou/core.py

@@ -10,6 +10,7 @@ from base.base_crawler import AbstractCrawler
 from base.proxy_account_pool import AccountPool
 from tools import utils
 from var import crawler_type_var
+from models import kuaishou
 
 from .client import KuaiShouClient
 from .login import KuaishouLogin
@@ -91,42 +92,38 @@ class KuaishouCrawler(AbstractCrawler):
                 if not videos_res:
                     utils.logger.error(f"search info by keyword:{keyword} not found data")
                     continue
-                vision_search_photo = videos_res.get("visionSearchPhoto")
-                utils.logger.info(f"videos_res:{videos_res}")
-                semaphore = asyncio.Semaphore(config.MAX_CONCURRENCY_NUM)
-                task_list = [
-                    self.get_video_detail(feed_item.get("photo", {}).get("id"), semaphore)
-                    for feed_item in vision_search_photo.get("feeds", {})
-                ]
-                video_details = await asyncio.gather(*task_list)
-                for video_detail in video_details:
-                    if video_detail is not None:
-                        video_id_list.append(video_detail.get("id"))
-                page += 1
-                utils.logger.info(f"Video details: {video_details}")
-                await self.batch_get_note_comments(video_id_list)
 
-        await asyncio.Event().wait()
+                vision_search_photo: Dict = videos_res.get("visionSearchPhoto")
+                if vision_search_photo.get("result") != 1:
+                    utils.logger.error(f"search info by keyword:{keyword} not found data ")
+                    continue
+
+                for video_detail in vision_search_photo.get("feeds"):
+                    video_id_list.append(video_detail.get("photo", {}).get("id"))
+                    await kuaishou.update_kuaishou_video(video_item=video_detail)
+
+                # batch fetch video comments
+                page += 1
+                await self.batch_get_video_comments(video_id_list)
 
     async def get_specified_notes(self):
         pass
 
-    async def batch_get_note_comments(self, video_id_list: List[str]):
-        pass
+    async def batch_get_video_comments(self, video_id_list: List[str]):
+        utils.logger.info(f"[batch_get_video_comments] video ids:{video_id_list}")
 
-    async def get_video_detail(self, ):
-        pass
-
-    async def get_video_detail(self, video_id: str, semaphore: asyncio.Semaphore) -> Optional[Dict]:
-        """Get note detail"""
+    async def get_video_info_task(self, video_id: str, semaphore: asyncio.Semaphore) -> Optional[Dict]:
+        """Get video detail task"""
         async with semaphore:
             try:
-                return await self.ks_client.get_video_info(video_id)
+                result = await self.ks_client.get_video_info(video_id)
+                utils.logger.info(f"Get video_id:{video_id} info result: {result} ...")
+                return result
             except DataFetchError as ex:
-                utils.logger.error(f"Get note detail error: {ex}")
+                utils.logger.error(f"Get video detail error: {ex}")
                 return None
             except KeyError as ex:
-                utils.logger.error(f"have not fund note detail note_id:{note_id}, err: {ex}")
+                utils.logger.error(f"have not fund note detail video_id:{video_id}, err: {ex}")
                 return None
 
     def create_proxy_info(self) -> Tuple[Optional[str], Optional[Dict], Optional[str]]:

media_platform/kuaishou/graphql/video_detail.graphql

@@ -0,0 +1,80 @@
+query visionVideoDetail($photoId: String, $type: String, $page: String, $webPageArea: String) {
+  visionVideoDetail(photoId: $photoId, type: $type, page: $page, webPageArea: $webPageArea) {
+    status
+    type
+    author {
+      id
+      name
+      following
+      headerUrl
+      __typename
+    }
+    photo {
+      id
+      duration
+      caption
+      likeCount
+      realLikeCount
+      coverUrl
+      photoUrl
+      liked
+      timestamp
+      expTag
+      llsid
+      viewCount
+      videoRatio
+      stereoType
+      musicBlocked
+      manifest {
+        mediaType
+        businessType
+        version
+        adaptationSet {
+          id
+          duration
+          representation {
+            id
+            defaultSelect
+            backupUrl
+            codecs
+            url
+            height
+            width
+            avgBitrate
+            maxBitrate
+            m3u8Slice
+            qualityType
+            qualityLabel
+            frameRate
+            featureP2sp
+            hidden
+            disableAdaptive
+            __typename
+          }
+          __typename
+        }
+        __typename
+      }
+      manifestH265
+      photoH265Url
+      coronaCropManifest
+      coronaCropManifestH265
+      croppedPhotoH265Url
+      croppedPhotoUrl
+      videoResource
+      __typename
+    }
+    tags {
+      type
+      name
+      __typename
+    }
+    commentLimit {
+      canAddComment
+      __typename
+    }
+    llsid
+    danmakuSwitch
+    __typename
+  }
+}

media_platform/kuaishou/login.py

@@ -28,7 +28,16 @@ class KuaishouLogin(AbstractLogin):
         self.cookie_str = cookie_str
 
     async def begin(self):
-        pass
+        """Start login xiaohongshu"""
+        utils.logger.info("Begin login kuaishou ...")
+        if self.login_type == "qrcode":
+            await self.login_by_qrcode()
+        elif self.login_type == "phone":
+            await self.login_by_mobile()
+        elif self.login_type == "cookie":
+            await self.login_by_cookies()
+        else:
+            raise ValueError("Invalid Login Type Currently only supported qrcode or phone or cookie ...")
 
     async def login_by_qrcode(self):
         pass
@@ -37,4 +46,11 @@ class KuaishouLogin(AbstractLogin):
         pass
 
     async def login_by_cookies(self):
-        pass
+        utils.logger.info("Begin login kuaishou by cookie ...")
+        for key, value in utils.convert_str_cookie_to_dict(self.cookie_str).items():
+            await self.browser_context.add_cookies([{
+                'name': key,
+                'value': value,
+                'domain': ".douyin.com",
+                'path': "/"
+            }])