mirror of
https://github.com/RYDE-WORK/MediaCrawler.git
synced 2026-01-19 13:33:27 +08:00
feat: douyin adds comment images
This commit is contained in:
Relakkes
parent
a8e796e690
commit
30d0e733d5
@ -536,4 +536,5 @@ CREATE TABLE `zhihu_creator` (
|
||||
-- add column `like_count` to douyin_aweme_comment
|
||||
alter table douyin_aweme_comment add column `like_count` varchar(255) NOT NULL DEFAULT '0' COMMENT '点赞数';
|
||||
|
||||
alter table xhs_note add column xsec_token varchar(50) default null comment '签名算法';
|
||||
alter table xhs_note add column xsec_token varchar(50) default null comment '签名算法';
|
||||
alter table douyin_aweme_comment add column `pictures` varchar(500) NOT NULL DEFAULT '' COMMENT '评论图片列表';
|
||||
@ -1,12 +1,12 @@
|
||||
# 声明：本代码仅供学习和研究目的使用。使用者应遵守以下原则：
|
||||
# 1. 不得用于任何商业用途。
|
||||
# 2. 使用时应遵守目标平台的使用条款和robots.txt规则。
|
||||
# 3. 不得进行大规模爬取或对平台造成运营干扰。
|
||||
# 4. 应合理控制请求频率，避免给目标平台带来不必要的负担。
|
||||
# 声明：本代码仅供学习和研究目的使用。使用者应遵守以下原则：
|
||||
# 1. 不得用于任何商业用途。
|
||||
# 2. 使用时应遵守目标平台的使用条款和robots.txt规则。
|
||||
# 3. 不得进行大规模爬取或对平台造成运营干扰。
|
||||
# 4. 应合理控制请求频率，避免给目标平台带来不必要的负担。
|
||||
# 5. 不得用于任何非法或不当的用途。
|
||||
#
|
||||
# 详细许可条款请参阅项目根目录下的LICENSE文件。
|
||||
# 使用本代码即表示您同意遵守上述原则和LICENSE中的所有条款。
|
||||
#
|
||||
# 详细许可条款请参阅项目根目录下的LICENSE文件。
|
||||
# 使用本代码即表示您同意遵守上述原则和LICENSE中的所有条款。
|
||||
|
||||
|
||||
# -*- coding: utf-8 -*-
|
||||
@ -25,7 +25,7 @@ class DouyinStoreFactory:
|
||||
STORES = {
|
||||
"csv": DouyinCsvStoreImplement,
|
||||
"db": DouyinDbStoreImplement,
|
||||
"json": DouyinJsonStoreImplement
|
||||
"json": DouyinJsonStoreImplement,
|
||||
}
|
||||
|
||||
@staticmethod
|
||||
@ -33,10 +33,35 @@ class DouyinStoreFactory:
|
||||
store_class = DouyinStoreFactory.STORES.get(config.SAVE_DATA_OPTION)
|
||||
if not store_class:
|
||||
raise ValueError(
|
||||
"[DouyinStoreFactory.create_store] Invalid save option only supported csv or db or json ...")
|
||||
"[DouyinStoreFactory.create_store] Invalid save option only supported csv or db or json ..."
|
||||
)
|
||||
return store_class()
|
||||
|
||||
|
||||
def _extract_comment_image_list(comment_item: Dict) -> List[str]:
|
||||
"""
|
||||
提取评论图片列表
|
||||
|
||||
Args:
|
||||
comment_item (Dict): 抖音评论
|
||||
|
||||
Returns:
|
||||
List[str]: 评论图片列表
|
||||
"""
|
||||
images_res: List[str] = []
|
||||
image_list: List[Dict] = comment_item.get("image_list", [])
|
||||
|
||||
if not image_list:
|
||||
return []
|
||||
|
||||
for image in image_list:
|
||||
image_url_list = image.get("origin_url", {}).get("url_list", [])
|
||||
if image_url_list and len(image_url_list) > 1:
|
||||
images_res.append(image_url_list[1])
|
||||
|
||||
return images_res
|
||||
|
||||
|
||||
async def update_douyin_aweme(aweme_item: Dict):
|
||||
aweme_id = aweme_item.get("aweme_id")
|
||||
user_info = aweme_item.get("author", {})
|
||||
@ -64,8 +89,11 @@ async def update_douyin_aweme(aweme_item: Dict):
|
||||
"source_keyword": source_keyword_var.get(),
|
||||
}
|
||||
utils.logger.info(
|
||||
f"[store.douyin.update_douyin_aweme] douyin aweme id:{aweme_id}, title:{save_content_item.get('title')}")
|
||||
await DouyinStoreFactory.create_store().store_content(content_item=save_content_item)
|
||||
f"[store.douyin.update_douyin_aweme] douyin aweme id:{aweme_id}, title:{save_content_item.get('title')}"
|
||||
)
|
||||
await DouyinStoreFactory.create_store().store_content(
|
||||
content_item=save_content_item
|
||||
)
|
||||
|
||||
|
||||
async def batch_update_dy_aweme_comments(aweme_id: str, comments: List[Dict]):
|
||||
@ -79,13 +107,19 @@ async def update_dy_aweme_comment(aweme_id: str, comment_item: Dict):
|
||||
comment_aweme_id = comment_item.get("aweme_id")
|
||||
if aweme_id != comment_aweme_id:
|
||||
utils.logger.error(
|
||||
f"[store.douyin.update_dy_aweme_comment] comment_aweme_id: {comment_aweme_id} != aweme_id: {aweme_id}")
|
||||
f"[store.douyin.update_dy_aweme_comment] comment_aweme_id: {comment_aweme_id} != aweme_id: {aweme_id}"
|
||||
)
|
||||
return
|
||||
user_info = comment_item.get("user", {})
|
||||
comment_id = comment_item.get("cid")
|
||||
parent_comment_id = comment_item.get("reply_id", "0")
|
||||
avatar_info = user_info.get("avatar_medium", {}) or user_info.get("avatar_300x300", {}) or user_info.get(
|
||||
"avatar_168x168", {}) or user_info.get("avatar_thumb", {}) or {}
|
||||
avatar_info = (
|
||||
user_info.get("avatar_medium", {})
|
||||
or user_info.get("avatar_300x300", {})
|
||||
or user_info.get("avatar_168x168", {})
|
||||
or user_info.get("avatar_thumb", {})
|
||||
or {}
|
||||
)
|
||||
save_comment_item = {
|
||||
"comment_id": comment_id,
|
||||
"create_time": comment_item.get("create_time"),
|
||||
@ -100,39 +134,39 @@ async def update_dy_aweme_comment(aweme_id: str, comment_item: Dict):
|
||||
"nickname": user_info.get("nickname"),
|
||||
"avatar": avatar_info.get("url_list", [""])[0],
|
||||
"sub_comment_count": str(comment_item.get("reply_comment_total", 0)),
|
||||
"like_count": comment_item.get("digg_count") if comment_item.get("digg_count") else 0,
|
||||
"like_count": (
|
||||
comment_item.get("digg_count") if comment_item.get("digg_count") else 0
|
||||
),
|
||||
"last_modify_ts": utils.get_current_timestamp(),
|
||||
"parent_comment_id": parent_comment_id
|
||||
"parent_comment_id": parent_comment_id,
|
||||
"pictures": ",".join(_extract_comment_image_list(comment_item)),
|
||||
}
|
||||
utils.logger.info(
|
||||
f"[store.douyin.update_dy_aweme_comment] douyin aweme comment: {comment_id}, content: {save_comment_item.get('content')}")
|
||||
|
||||
await DouyinStoreFactory.create_store().store_comment(comment_item=save_comment_item)
|
||||
|
||||
f"[store.douyin.update_dy_aweme_comment] douyin aweme comment: {comment_id}, content: {save_comment_item.get('content')}"
|
||||
)
|
||||
|
||||
await DouyinStoreFactory.create_store().store_comment(
|
||||
comment_item=save_comment_item
|
||||
)
|
||||
|
||||
|
||||
async def save_creator(user_id: str, creator: Dict):
|
||||
user_info = creator.get('user', {})
|
||||
gender_map = {
|
||||
0: '未知',
|
||||
1: '男',
|
||||
2: '女'
|
||||
}
|
||||
avatar_uri = user_info.get('avatar_300x300', {}).get('uri')
|
||||
user_info = creator.get("user", {})
|
||||
gender_map = {0: "未知", 1: "男", 2: "女"}
|
||||
avatar_uri = user_info.get("avatar_300x300", {}).get("uri")
|
||||
local_db_item = {
|
||||
'user_id': user_id,
|
||||
'nickname': user_info.get('nickname'),
|
||||
'gender': gender_map.get(user_info.get('gender'), '未知'),
|
||||
'avatar': f"https://p3-pc.douyinpic.com/img/{avatar_uri}" + r"~c5_300x300.jpeg?from=2956013662",
|
||||
'desc': user_info.get('signature'),
|
||||
'ip_location': user_info.get('ip_location'),
|
||||
'follows': user_info.get("following_count", 0),
|
||||
'fans': user_info.get("max_follower_count", 0),
|
||||
'interaction': user_info.get("total_favorited", 0),
|
||||
'videos_count': user_info.get("aweme_count", 0),
|
||||
"user_id": user_id,
|
||||
"nickname": user_info.get("nickname"),
|
||||
"gender": gender_map.get(user_info.get("gender"), "未知"),
|
||||
"avatar": f"https://p3-pc.douyinpic.com/img/{avatar_uri}"
|
||||
+ r"~c5_300x300.jpeg?from=2956013662",
|
||||
"desc": user_info.get("signature"),
|
||||
"ip_location": user_info.get("ip_location"),
|
||||
"follows": user_info.get("following_count", 0),
|
||||
"fans": user_info.get("max_follower_count", 0),
|
||||
"interaction": user_info.get("total_favorited", 0),
|
||||
"videos_count": user_info.get("aweme_count", 0),
|
||||
"last_modify_ts": utils.get_current_timestamp(),
|
||||
|
||||
}
|
||||
utils.logger.info(f"[store.douyin.save_creator] creator:{local_db_item}")
|
||||
await DouyinStoreFactory.create_store().store_creator(local_db_item)
|
||||
|
||||
Loading…
x
Reference in New Issue
Block a user