|
@@ -215,7 +215,7 @@ class GZXHAuthor:
|
|
|
|
|
|
|
|
def process_video_obj(self, article, user_name, wechat_gh):
|
|
def process_video_obj(self, article, user_name, wechat_gh):
|
|
|
trace_id = self.platform + str(uuid.uuid1())
|
|
trace_id = self.platform + str(uuid.uuid1())
|
|
|
- update_time_stamp = int(time.time())
|
|
|
|
|
|
|
+ # update_time_stamp = int(time.time())
|
|
|
publish_time_str = article.get("published_time", 0)
|
|
publish_time_str = article.get("published_time", 0)
|
|
|
date_format = "%Y-%m-%d %H:%M:%S"
|
|
date_format = "%Y-%m-%d %H:%M:%S"
|
|
|
date_time_obj = datetime.strptime(publish_time_str, date_format)
|
|
date_time_obj = datetime.strptime(publish_time_str, date_format)
|
|
@@ -233,7 +233,7 @@ class GZXHAuthor:
|
|
|
.replace("'", ""),
|
|
.replace("'", ""),
|
|
|
"out_video_id": video_id,
|
|
"out_video_id": video_id,
|
|
|
"publish_time_stamp": publish_time_stamp,
|
|
"publish_time_stamp": publish_time_stamp,
|
|
|
- "update_time_stamp": update_time_stamp,
|
|
|
|
|
|
|
+ "update_time_stamp": 0,
|
|
|
"publish_time_str": publish_time_str,
|
|
"publish_time_str": publish_time_str,
|
|
|
"play_cnt": 0,
|
|
"play_cnt": 0,
|
|
|
"comment_cnt": 0,
|
|
"comment_cnt": 0,
|
|
@@ -245,7 +245,7 @@ class GZXHAuthor:
|
|
|
"width": 0,
|
|
"width": 0,
|
|
|
"height": 0,
|
|
"height": 0,
|
|
|
"duration": 0,
|
|
"duration": 0,
|
|
|
- "platform": 'gongzhonghao',
|
|
|
|
|
|
|
+ "platform": self.platform,
|
|
|
"strategy": self.mode,
|
|
"strategy": self.mode,
|
|
|
"crawler_rule": self.rule_dict,
|
|
"crawler_rule": self.rule_dict,
|
|
|
"session": f"gongzhongxinhao-author-{int(time.time())}",
|
|
"session": f"gongzhongxinhao-author-{int(time.time())}",
|
|
@@ -261,21 +261,21 @@ class GZXHAuthor:
|
|
|
data=video_dict,
|
|
data=video_dict,
|
|
|
)
|
|
)
|
|
|
pipeline = PiaoQuanPipeline(
|
|
pipeline = PiaoQuanPipeline(
|
|
|
- platform='gongzhonghao',
|
|
|
|
|
|
|
+ platform=self.platform,
|
|
|
mode=self.mode,
|
|
mode=self.mode,
|
|
|
item=video_dict,
|
|
item=video_dict,
|
|
|
rule_dict=self.rule_dict,
|
|
rule_dict=self.rule_dict,
|
|
|
env=self.env,
|
|
env=self.env,
|
|
|
trace_id=trace_id,
|
|
trace_id=trace_id,
|
|
|
)
|
|
)
|
|
|
|
|
+ if not pipeline.publish_time_flag():
|
|
|
|
|
+ return False
|
|
|
if not pipeline.repeat_video():
|
|
if not pipeline.repeat_video():
|
|
|
return True
|
|
return True
|
|
|
if not pipeline.download_rule_flag():
|
|
if not pipeline.download_rule_flag():
|
|
|
return True
|
|
return True
|
|
|
if not pipeline.title_flag():
|
|
if not pipeline.title_flag():
|
|
|
return True
|
|
return True
|
|
|
- if not pipeline.publish_time_flag():
|
|
|
|
|
- return True
|
|
|
|
|
else:
|
|
else:
|
|
|
current_time = datetime.now()
|
|
current_time = datetime.now()
|
|
|
timestamp = current_time.strftime("%Y-%m-%d %H:%M:%S")
|
|
timestamp = current_time.strftime("%Y-%m-%d %H:%M:%S")
|
|
@@ -317,7 +317,7 @@ if __name__ == "__main__":
|
|
|
platform="gongzhonghao",
|
|
platform="gongzhonghao",
|
|
|
mode="author",
|
|
mode="author",
|
|
|
user_dict={"uid": "123456", "link": "https://mp.weixin.qq.com/s/Cwc2D3RUNDk30zv_s0IxTA", "user_id": "1234565"},
|
|
user_dict={"uid": "123456", "link": "https://mp.weixin.qq.com/s/Cwc2D3RUNDk30zv_s0IxTA", "user_id": "1234565"},
|
|
|
- rule_dict={"rule": "{period:{min:1,max:1}}"},
|
|
|
|
|
|
|
+ rule_dict={"period":{"min":1,"max":1}},
|
|
|
env="dev",
|
|
env="dev",
|
|
|
)
|
|
)
|
|
|
|
|
|
