refactor: 规范日志打印

feat: B站指定视频ID爬取(bvid)
This commit is contained in:
Relakkes
2023-12-23 01:04:08 +08:00
parent 273c9a316b
commit aba9f14f50
18 changed files with 147 additions and 133 deletions

View File

@@ -59,12 +59,12 @@ class KuaiShouClient:
@staticmethod
async def pong() -> bool:
"""get a note to check if login state is ok"""
utils.logger.info("Begin pong kuaishou...")
utils.logger.info("[KuaiShouClient.pong] Begin pong kuaishou...")
ping_flag = False
try:
pass
except Exception as e:
utils.logger.error(f"Pong kuaishou failed: {e}, and try to login again...")
utils.logger.error(f"[KuaiShouClient.pong] Pong kuaishou failed: {e}, and try to login again...")
ping_flag = False
return ping_flag

View File

@@ -81,13 +81,13 @@ class KuaishouCrawler(AbstractCrawler):
else:
pass
utils.logger.info("Kuaishou Crawler finished ...")
utils.logger.info("[KuaishouCrawler.start] Kuaishou Crawler finished ...")
async def search(self):
utils.logger.info("Begin search kuaishou keywords")
utils.logger.info("[KuaishouCrawler.search] Begin search kuaishou keywords")
ks_limit_count = 20 # kuaishou limit page fixed value
for keyword in config.KEYWORDS.split(","):
utils.logger.info(f"Current search keyword: {keyword}")
utils.logger.info(f"[KuaishouCrawler.search] Current search keyword: {keyword}")
page = 1
while page * ks_limit_count <= config.CRAWLER_MAX_NOTES_COUNT:
video_id_list: List[str] = []
@@ -96,12 +96,12 @@ class KuaishouCrawler(AbstractCrawler):
pcursor=str(page),
)
if not videos_res:
utils.logger.error(f"search info by keyword:{keyword} not found data")
utils.logger.error(f"[KuaishouCrawler.search] search info by keyword:{keyword} not found data")
continue
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 ")
utils.logger.error(f"[KuaishouCrawler.search] search info by keyword:{keyword} not found data ")
continue
for video_detail in vision_search_photo.get("feeds"):
@@ -129,13 +129,13 @@ class KuaishouCrawler(AbstractCrawler):
async with semaphore:
try:
result = await self.ks_client.get_video_info(video_id)
utils.logger.info(f"Get video_id:{video_id} info result: {result} ...")
utils.logger.info(f"[KuaishouCrawler.get_video_info_task] Get video_id:{video_id} info result: {result} ...")
return result.get("visionVideoDetail")
except DataFetchError as ex:
utils.logger.error(f"Get video detail error: {ex}")
utils.logger.error(f"[KuaishouCrawler.get_video_info_task] Get video detail error: {ex}")
return None
except KeyError as ex:
utils.logger.error(f"have not fund note detail video_id:{video_id}, err: {ex}")
utils.logger.error(f"[KuaishouCrawler.get_video_info_task] have not fund note detail video_id:{video_id}, err: {ex}")
return None
async def batch_get_video_comments(self, video_id_list: List[str]):
@@ -144,7 +144,7 @@ class KuaishouCrawler(AbstractCrawler):
:param video_id_list:
:return:
"""
utils.logger.info(f"[batch_get_video_comments] video ids:{video_id_list}")
utils.logger.info(f"[KuaishouCrawler.batch_get_video_comments] video ids:{video_id_list}")
semaphore = asyncio.Semaphore(config.MAX_CONCURRENCY_NUM)
task_list: List[Task] = []
for video_id in video_id_list:
@@ -163,16 +163,16 @@ class KuaishouCrawler(AbstractCrawler):
"""
async with semaphore:
try:
utils.logger.info(f"[get_comments] bengin get video_id: {video_id} comments ...")
utils.logger.info(f"[KuaishouCrawler.get_comments] begin get video_id: {video_id} comments ...")
await self.ks_client.get_video_all_comments(
photo_id=video_id,
crawl_interval=random.random(),
callback=kuaishou.batch_update_ks_video_comments
)
except DataFetchError as ex:
utils.logger.error(f"[get_comments] get video_id: {video_id} comment error: {ex}")
utils.logger.error(f"[KuaishouCrawler.get_comments] get video_id: {video_id} comment error: {ex}")
except Exception as e:
utils.logger.error(f"[get_comments] may be been blocked, err:{e}")
utils.logger.error(f"[KuaishouCrawler.get_comments] may be been blocked, err:{e}")
# use time.sleeep block main coroutine instead of asyncio.sleep and cacel running comment task
# maybe kuaishou block our request, we will take a nap and update the cookie again
current_running_tasks = comment_tasks_var.get()
@@ -197,7 +197,7 @@ class KuaishouCrawler(AbstractCrawler):
async def create_ks_client(self, httpx_proxy: Optional[str]) -> KuaiShouClient:
"""Create xhs client"""
utils.logger.info("Begin create kuaishou API client ...")
utils.logger.info("[KuaishouCrawler.create_ks_client] Begin create kuaishou API client ...")
cookie_str, cookie_dict = utils.convert_cookies(await self.browser_context.cookies())
xhs_client_obj = KuaiShouClient(
proxies=httpx_proxy,
@@ -221,7 +221,7 @@ class KuaishouCrawler(AbstractCrawler):
headless: bool = True
) -> BrowserContext:
"""Launch browser and create browser context"""
utils.logger.info("Begin create browser context ...")
utils.logger.info("[KuaishouCrawler.launch_browser] Begin create browser context ...")
if config.SAVE_LOGIN_STATE:
user_data_dir = os.path.join(os.getcwd(), "browser_data",
config.USER_DATA_DIR % self.platform) # type: ignore
@@ -245,4 +245,4 @@ class KuaishouCrawler(AbstractCrawler):
async def close(self):
"""Close browser context"""
await self.browser_context.close()
utils.logger.info("Browser context closed ...")
utils.logger.info("[KuaishouCrawler.close] Browser context closed ...")

View File

@@ -29,7 +29,7 @@ class KuaishouLogin(AbstractLogin):
async def begin(self):
"""Start login xiaohongshu"""
utils.logger.info("Begin login kuaishou ...")
utils.logger.info("[KuaishouLogin.begin] Begin login kuaishou ...")
if self.login_type == "qrcode":
await self.login_by_qrcode()
elif self.login_type == "phone":
@@ -37,7 +37,7 @@ class KuaishouLogin(AbstractLogin):
elif self.login_type == "cookie":
await self.login_by_cookies()
else:
raise ValueError("Invalid Login Type Currently only supported qrcode or phone or cookie ...")
raise ValueError("[KuaishouLogin.begin] Invalid Login Type Currently only supported qrcode or phone or cookie ...")
@retry(stop=stop_after_attempt(20), wait=wait_fixed(1), retry=retry_if_result(lambda value: value is False))
async def check_login_state(self) -> bool:
@@ -55,7 +55,7 @@ class KuaishouLogin(AbstractLogin):
async def login_by_qrcode(self):
"""login kuaishou website and keep webdriver login state"""
utils.logger.info("Begin login kuaishou by qrcode ...")
utils.logger.info("[KuaishouLogin.login_by_qrcode] Begin login kuaishou by qrcode ...")
# click login button
login_button_ele = self.context_page.locator(
@@ -70,7 +70,7 @@ class KuaishouLogin(AbstractLogin):
selector=qrcode_img_selector
)
if not base64_qrcode_img:
utils.logger.info("login failed , have not found qrcode please check ....")
utils.logger.info("[KuaishouLogin.login_by_qrcode] login failed , have not found qrcode please check ....")
sys.exit()
@@ -78,22 +78,22 @@ class KuaishouLogin(AbstractLogin):
partial_show_qrcode = functools.partial(utils.show_qrcode, base64_qrcode_img)
asyncio.get_running_loop().run_in_executor(executor=None, func=partial_show_qrcode)
utils.logger.info(f"waiting for scan code login, remaining time is 20s")
utils.logger.info(f"[KuaishouLogin.login_by_qrcode] waiting for scan code login, remaining time is 20s")
try:
await self.check_login_state()
except RetryError:
utils.logger.info("Login kuaishou failed by qrcode login method ...")
utils.logger.info("[KuaishouLogin.login_by_qrcode] Login kuaishou failed by qrcode login method ...")
sys.exit()
wait_redirect_seconds = 5
utils.logger.info(f"Login successful then wait for {wait_redirect_seconds} seconds redirect ...")
utils.logger.info(f"[KuaishouLogin.login_by_qrcode] Login successful then wait for {wait_redirect_seconds} seconds redirect ...")
await asyncio.sleep(wait_redirect_seconds)
async def login_by_mobile(self):
pass
async def login_by_cookies(self):
utils.logger.info("Begin login kuaishou by cookie ...")
utils.logger.info("[KuaishouLogin.login_by_cookies] 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,