Uploading the AI Crawler System: MindSpider
This commit is contained in:
戒酒的李白
@@ -0,0 +1,220 @@
|
||||
# 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。
|
||||
|
||||
# -*- coding: utf-8 -*-
|
||||
# @Author : relakkes@gmail.com
|
||||
# @Time : 2024/1/14 19:34
|
||||
# @Desc :
|
||||
|
||||
from typing import List
|
||||
|
||||
import config
|
||||
from var import source_keyword_var
|
||||
|
||||
from .bilibili_store_impl import *
|
||||
from .bilibilli_store_media import *
|
||||
|
||||
|
||||
class BiliStoreFactory:
|
||||
STORES = {
|
||||
"csv": BiliCsvStoreImplement,
|
||||
"db": BiliDbStoreImplement,
|
||||
"json": BiliJsonStoreImplement,
|
||||
"sqlite": BiliSqliteStoreImplement,
|
||||
}
|
||||
|
||||
@staticmethod
|
||||
def create_store() -> AbstractStore:
|
||||
store_class = BiliStoreFactory.STORES.get(config.SAVE_DATA_OPTION)
|
||||
if not store_class:
|
||||
raise ValueError("[BiliStoreFactory.create_store] Invalid save option only supported csv or db or json or sqlite ...")
|
||||
return store_class()
|
||||
|
||||
|
||||
async def update_bilibili_video(video_item: Dict):
|
||||
video_item_view: Dict = video_item.get("View")
|
||||
video_user_info: Dict = video_item_view.get("owner")
|
||||
video_item_stat: Dict = video_item_view.get("stat")
|
||||
video_id = str(video_item_view.get("aid"))
|
||||
save_content_item = {
|
||||
"video_id": video_id,
|
||||
"video_type": "video",
|
||||
"title": video_item_view.get("title", "")[:500],
|
||||
"desc": video_item_view.get("desc", "")[:500],
|
||||
"create_time": video_item_view.get("pubdate"),
|
||||
"user_id": str(video_user_info.get("mid")),
|
||||
"nickname": video_user_info.get("name"),
|
||||
"avatar": video_user_info.get("face", ""),
|
||||
"liked_count": str(video_item_stat.get("like", "")),
|
||||
"disliked_count": str(video_item_stat.get("dislike", "")),
|
||||
"video_play_count": str(video_item_stat.get("view", "")),
|
||||
"video_favorite_count": str(video_item_stat.get("favorite", "")),
|
||||
"video_share_count": str(video_item_stat.get("share", "")),
|
||||
"video_coin_count": str(video_item_stat.get("coin", "")),
|
||||
"video_danmaku": str(video_item_stat.get("danmaku", "")),
|
||||
"video_comment": str(video_item_stat.get("reply", "")),
|
||||
"last_modify_ts": utils.get_current_timestamp(),
|
||||
"video_url": f"https://www.bilibili.com/video/av{video_id}",
|
||||
"video_cover_url": video_item_view.get("pic", ""),
|
||||
"source_keyword": source_keyword_var.get(),
|
||||
}
|
||||
utils.logger.info(f"[store.bilibili.update_bilibili_video] bilibili video id:{video_id}, title:{save_content_item.get('title')}")
|
||||
await BiliStoreFactory.create_store().store_content(content_item=save_content_item)
|
||||
|
||||
|
||||
async def update_up_info(video_item: Dict):
|
||||
video_item_card_list: Dict = video_item.get("Card")
|
||||
video_item_card: Dict = video_item_card_list.get("card")
|
||||
saver_up_info = {
|
||||
"user_id": str(video_item_card.get("mid")),
|
||||
"nickname": video_item_card.get("name"),
|
||||
"sex": video_item_card.get("sex"),
|
||||
"sign": video_item_card.get("sign"),
|
||||
"avatar": video_item_card.get("face"),
|
||||
"last_modify_ts": utils.get_current_timestamp(),
|
||||
"total_fans": video_item_card.get("fans"),
|
||||
"total_liked": video_item_card_list.get("like_num"),
|
||||
"user_rank": video_item_card.get("level_info").get("current_level"),
|
||||
"is_official": video_item_card.get("official_verify").get("type"),
|
||||
}
|
||||
utils.logger.info(f"[store.bilibili.update_up_info] bilibili user_id:{video_item_card.get('mid')}")
|
||||
await BiliStoreFactory.create_store().store_creator(creator=saver_up_info)
|
||||
|
||||
|
||||
async def batch_update_bilibili_video_comments(video_id: str, comments: List[Dict]):
|
||||
if not comments:
|
||||
return
|
||||
for comment_item in comments:
|
||||
await update_bilibili_video_comment(video_id, comment_item)
|
||||
|
||||
|
||||
async def update_bilibili_video_comment(video_id: str, comment_item: Dict):
|
||||
comment_id = str(comment_item.get("rpid"))
|
||||
parent_comment_id = str(comment_item.get("parent", 0))
|
||||
content: Dict = comment_item.get("content")
|
||||
user_info: Dict = comment_item.get("member")
|
||||
like_count: int = comment_item.get("like", 0)
|
||||
save_comment_item = {
|
||||
"comment_id": comment_id,
|
||||
"parent_comment_id": parent_comment_id,
|
||||
"create_time": comment_item.get("ctime"),
|
||||
"video_id": str(video_id),
|
||||
"content": content.get("message"),
|
||||
"user_id": user_info.get("mid"),
|
||||
"nickname": user_info.get("uname"),
|
||||
"sex": user_info.get("sex"),
|
||||
"sign": user_info.get("sign"),
|
||||
"avatar": user_info.get("avatar"),
|
||||
"sub_comment_count": str(comment_item.get("rcount", 0)),
|
||||
"like_count": like_count,
|
||||
"last_modify_ts": utils.get_current_timestamp(),
|
||||
}
|
||||
utils.logger.info(f"[store.bilibili.update_bilibili_video_comment] Bilibili video comment: {comment_id}, content: {save_comment_item.get('content')}")
|
||||
await BiliStoreFactory.create_store().store_comment(comment_item=save_comment_item)
|
||||
|
||||
|
||||
async def store_video(aid, video_content, extension_file_name):
|
||||
"""
|
||||
video video storage implementation
|
||||
Args:
|
||||
aid:
|
||||
video_content:
|
||||
extension_file_name:
|
||||
"""
|
||||
await BilibiliVideo().store_video({
|
||||
"aid": aid,
|
||||
"video_content": video_content,
|
||||
"extension_file_name": extension_file_name,
|
||||
})
|
||||
|
||||
|
||||
async def batch_update_bilibili_creator_fans(creator_info: Dict, fans_list: List[Dict]):
|
||||
if not fans_list:
|
||||
return
|
||||
for fan_item in fans_list:
|
||||
fan_info: Dict = {
|
||||
"id": fan_item.get("mid"),
|
||||
"name": fan_item.get("uname"),
|
||||
"sign": fan_item.get("sign"),
|
||||
"avatar": fan_item.get("face"),
|
||||
}
|
||||
await update_bilibili_creator_contact(creator_info=creator_info, fan_info=fan_info)
|
||||
|
||||
|
||||
async def batch_update_bilibili_creator_followings(creator_info: Dict, followings_list: List[Dict]):
|
||||
if not followings_list:
|
||||
return
|
||||
for following_item in followings_list:
|
||||
following_info: Dict = {
|
||||
"id": following_item.get("mid"),
|
||||
"name": following_item.get("uname"),
|
||||
"sign": following_item.get("sign"),
|
||||
"avatar": following_item.get("face"),
|
||||
}
|
||||
await update_bilibili_creator_contact(creator_info=following_info, fan_info=creator_info)
|
||||
|
||||
|
||||
async def batch_update_bilibili_creator_dynamics(creator_info: Dict, dynamics_list: List[Dict]):
|
||||
if not dynamics_list:
|
||||
return
|
||||
for dynamic_item in dynamics_list:
|
||||
dynamic_id: str = dynamic_item["id_str"]
|
||||
dynamic_text: str = ""
|
||||
if dynamic_item["modules"]["module_dynamic"].get("desc"):
|
||||
dynamic_text = dynamic_item["modules"]["module_dynamic"]["desc"]["text"]
|
||||
dynamic_type: str = dynamic_item["type"].split("_")[-1]
|
||||
dynamic_pub_ts: str = dynamic_item["modules"]["module_author"]["pub_ts"]
|
||||
dynamic_stat: Dict = dynamic_item["modules"]["module_stat"]
|
||||
dynamic_comment: int = dynamic_stat["comment"]["count"]
|
||||
dynamic_forward: int = dynamic_stat["forward"]["count"]
|
||||
dynamic_like: int = dynamic_stat["like"]["count"]
|
||||
dynamic_info: Dict = {
|
||||
"dynamic_id": dynamic_id,
|
||||
"text": dynamic_text,
|
||||
"type": dynamic_type,
|
||||
"pub_ts": dynamic_pub_ts,
|
||||
"total_comments": dynamic_comment,
|
||||
"total_forwards": dynamic_forward,
|
||||
"total_liked": dynamic_like,
|
||||
}
|
||||
await update_bilibili_creator_dynamic(creator_info=creator_info, dynamic_info=dynamic_info)
|
||||
|
||||
|
||||
async def update_bilibili_creator_contact(creator_info: Dict, fan_info: Dict):
|
||||
save_contact_item = {
|
||||
"up_id": creator_info["id"],
|
||||
"fan_id": fan_info["id"],
|
||||
"up_name": creator_info["name"],
|
||||
"fan_name": fan_info["name"],
|
||||
"up_sign": creator_info["sign"],
|
||||
"fan_sign": fan_info["sign"],
|
||||
"up_avatar": creator_info["avatar"],
|
||||
"fan_avatar": fan_info["avatar"],
|
||||
"last_modify_ts": utils.get_current_timestamp(),
|
||||
}
|
||||
|
||||
await BiliStoreFactory.create_store().store_contact(contact_item=save_contact_item)
|
||||
|
||||
|
||||
async def update_bilibili_creator_dynamic(creator_info: Dict, dynamic_info: Dict):
|
||||
save_dynamic_item = {
|
||||
"dynamic_id": dynamic_info["dynamic_id"],
|
||||
"user_id": creator_info["id"],
|
||||
"user_name": creator_info["name"],
|
||||
"text": dynamic_info["text"],
|
||||
"type": dynamic_info["type"],
|
||||
"pub_ts": dynamic_info["pub_ts"],
|
||||
"total_comments": dynamic_info["total_comments"],
|
||||
"total_forwards": dynamic_info["total_forwards"],
|
||||
"total_liked": dynamic_info["total_liked"],
|
||||
"last_modify_ts": utils.get_current_timestamp(),
|
||||
}
|
||||
|
||||
await BiliStoreFactory.create_store().store_dynamic(dynamic_item=save_dynamic_item)
|
||||
@@ -0,0 +1,465 @@
|
||||
# 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。
|
||||
|
||||
|
||||
# -*- coding: utf-8 -*-
|
||||
# @Author : relakkes@gmail.com
|
||||
# @Time : 2024/1/14 19:34
|
||||
# @Desc : B站存储实现类
|
||||
import asyncio
|
||||
import csv
|
||||
import json
|
||||
import os
|
||||
import pathlib
|
||||
from typing import Dict
|
||||
|
||||
import aiofiles
|
||||
|
||||
import config
|
||||
from base.base_crawler import AbstractStore
|
||||
from tools import utils, words
|
||||
from var import crawler_type_var
|
||||
|
||||
|
||||
def calculate_number_of_files(file_store_path: str) -> int:
|
||||
"""计算数据保存文件的前部分排序数字，支持每次运行代码不写到同一个文件中
|
||||
Args:
|
||||
file_store_path;
|
||||
Returns:
|
||||
file nums
|
||||
"""
|
||||
if not os.path.exists(file_store_path):
|
||||
return 1
|
||||
try:
|
||||
return max([int(file_name.split("_")[0])for file_name in os.listdir(file_store_path)])+1
|
||||
except ValueError:
|
||||
return 1
|
||||
|
||||
class BiliCsvStoreImplement(AbstractStore):
|
||||
csv_store_path: str = "data/bilibili"
|
||||
file_count:int=calculate_number_of_files(csv_store_path)
|
||||
def make_save_file_name(self, store_type: str) -> str:
|
||||
"""
|
||||
make save file name by store type
|
||||
Args:
|
||||
store_type: contents or comments
|
||||
|
||||
Returns: eg: data/bilibili/search_comments_20240114.csv ...
|
||||
|
||||
"""
|
||||
return f"{self.csv_store_path}/{self.file_count}_{crawler_type_var.get()}_{store_type}_{utils.get_current_date()}.csv"
|
||||
|
||||
async def save_data_to_csv(self, save_item: Dict, store_type: str):
|
||||
"""
|
||||
Below is a simple way to save it in CSV format.
|
||||
Args:
|
||||
save_item: save content dict info
|
||||
store_type: Save type contains content and comments（contents | comments）
|
||||
|
||||
Returns: no returns
|
||||
|
||||
"""
|
||||
pathlib.Path(self.csv_store_path).mkdir(parents=True, exist_ok=True)
|
||||
save_file_name = self.make_save_file_name(store_type=store_type)
|
||||
async with aiofiles.open(save_file_name, mode='a+', encoding="utf-8-sig", newline="") as f:
|
||||
writer = csv.writer(f)
|
||||
if await f.tell() == 0:
|
||||
await writer.writerow(save_item.keys())
|
||||
await writer.writerow(save_item.values())
|
||||
|
||||
async def store_content(self, content_item: Dict):
|
||||
"""
|
||||
Bilibili content CSV storage implementation
|
||||
Args:
|
||||
content_item: note item dict
|
||||
|
||||
Returns:
|
||||
|
||||
"""
|
||||
await self.save_data_to_csv(save_item=content_item, store_type="contents")
|
||||
|
||||
async def store_comment(self, comment_item: Dict):
|
||||
"""
|
||||
Bilibili comment CSV storage implementation
|
||||
Args:
|
||||
comment_item: comment item dict
|
||||
|
||||
Returns:
|
||||
|
||||
"""
|
||||
await self.save_data_to_csv(save_item=comment_item, store_type="comments")
|
||||
|
||||
async def store_creator(self, creator: Dict):
|
||||
"""
|
||||
Bilibili creator CSV storage implementation
|
||||
Args:
|
||||
creator: creator item dict
|
||||
|
||||
Returns:
|
||||
|
||||
"""
|
||||
await self.save_data_to_csv(save_item=creator, store_type="creators")
|
||||
|
||||
async def store_contact(self, contact_item: Dict):
|
||||
"""
|
||||
Bilibili contact CSV storage implementation
|
||||
Args:
|
||||
contact_item: creator's contact item dict
|
||||
|
||||
Returns:
|
||||
|
||||
"""
|
||||
|
||||
await self.save_data_to_csv(save_item=contact_item, store_type="contacts")
|
||||
|
||||
async def store_dynamic(self, dynamic_item: Dict):
|
||||
"""
|
||||
Bilibili dynamic CSV storage implementation
|
||||
Args:
|
||||
dynamic_item: creator's dynamic item dict
|
||||
|
||||
Returns:
|
||||
|
||||
"""
|
||||
|
||||
await self.save_data_to_csv(save_item=dynamic_item, store_type="dynamics")
|
||||
|
||||
|
||||
class BiliDbStoreImplement(AbstractStore):
|
||||
async def store_content(self, content_item: Dict):
|
||||
"""
|
||||
Bilibili content DB storage implementation
|
||||
Args:
|
||||
content_item: content item dict
|
||||
|
||||
Returns:
|
||||
|
||||
"""
|
||||
|
||||
from .bilibili_store_sql import (add_new_content,
|
||||
query_content_by_content_id,
|
||||
update_content_by_content_id)
|
||||
video_id = content_item.get("video_id")
|
||||
video_detail: Dict = await query_content_by_content_id(content_id=video_id)
|
||||
if not video_detail:
|
||||
content_item["add_ts"] = utils.get_current_timestamp()
|
||||
await add_new_content(content_item)
|
||||
else:
|
||||
await update_content_by_content_id(video_id, content_item=content_item)
|
||||
|
||||
async def store_comment(self, comment_item: Dict):
|
||||
"""
|
||||
Bilibili content DB storage implementation
|
||||
Args:
|
||||
comment_item: comment item dict
|
||||
|
||||
Returns:
|
||||
|
||||
"""
|
||||
|
||||
from .bilibili_store_sql import (add_new_comment,
|
||||
query_comment_by_comment_id,
|
||||
update_comment_by_comment_id)
|
||||
comment_id = comment_item.get("comment_id")
|
||||
comment_detail: Dict = await query_comment_by_comment_id(comment_id=comment_id)
|
||||
if not comment_detail:
|
||||
comment_item["add_ts"] = utils.get_current_timestamp()
|
||||
await add_new_comment(comment_item)
|
||||
else:
|
||||
await update_comment_by_comment_id(comment_id, comment_item=comment_item)
|
||||
|
||||
async def store_creator(self, creator: Dict):
|
||||
"""
|
||||
Bilibili creator DB storage implementation
|
||||
Args:
|
||||
creator: creator item dict
|
||||
|
||||
Returns:
|
||||
|
||||
"""
|
||||
|
||||
from .bilibili_store_sql import (add_new_creator,
|
||||
query_creator_by_creator_id,
|
||||
update_creator_by_creator_id)
|
||||
creator_id = creator.get("user_id")
|
||||
creator_detail: Dict = await query_creator_by_creator_id(creator_id=creator_id)
|
||||
if not creator_detail:
|
||||
creator["add_ts"] = utils.get_current_timestamp()
|
||||
await add_new_creator(creator)
|
||||
else:
|
||||
await update_creator_by_creator_id(creator_id,creator_item=creator)
|
||||
|
||||
async def store_contact(self, contact_item: Dict):
|
||||
"""
|
||||
Bilibili contact DB storage implementation
|
||||
Args:
|
||||
contact_item: contact item dict
|
||||
|
||||
Returns:
|
||||
|
||||
"""
|
||||
|
||||
from .bilibili_store_sql import (add_new_contact,
|
||||
query_contact_by_up_and_fan,
|
||||
update_contact_by_id, )
|
||||
|
||||
up_id = contact_item.get("up_id")
|
||||
fan_id = contact_item.get("fan_id")
|
||||
contact_detail: Dict = await query_contact_by_up_and_fan(up_id=up_id, fan_id=fan_id)
|
||||
if not contact_detail:
|
||||
contact_item["add_ts"] = utils.get_current_timestamp()
|
||||
await add_new_contact(contact_item)
|
||||
else:
|
||||
key_id = contact_detail.get("id")
|
||||
await update_contact_by_id(id=key_id, contact_item=contact_item)
|
||||
|
||||
async def store_dynamic(self, dynamic_item):
|
||||
"""
|
||||
Bilibili dynamic DB storage implementation
|
||||
Args:
|
||||
dynamic_item: dynamic item dict
|
||||
|
||||
Returns:
|
||||
|
||||
"""
|
||||
|
||||
from .bilibili_store_sql import (add_new_dynamic,
|
||||
query_dynamic_by_dynamic_id,
|
||||
update_dynamic_by_dynamic_id)
|
||||
|
||||
dynamic_id = dynamic_item.get("dynamic_id")
|
||||
dynamic_detail = await query_dynamic_by_dynamic_id(dynamic_id=dynamic_id)
|
||||
if not dynamic_detail:
|
||||
dynamic_item["add_ts"] = utils.get_current_timestamp()
|
||||
await add_new_dynamic(dynamic_item)
|
||||
else:
|
||||
await update_dynamic_by_dynamic_id(dynamic_id, dynamic_item=dynamic_item)
|
||||
|
||||
|
||||
class BiliJsonStoreImplement(AbstractStore):
|
||||
json_store_path: str = "data/bilibili/json"
|
||||
words_store_path: str = "data/bilibili/words"
|
||||
lock = asyncio.Lock()
|
||||
file_count:int=calculate_number_of_files(json_store_path)
|
||||
WordCloud = words.AsyncWordCloudGenerator()
|
||||
|
||||
|
||||
def make_save_file_name(self, store_type: str) -> (str,str):
|
||||
"""
|
||||
make save file name by store type
|
||||
Args:
|
||||
store_type: Save type contains content and comments（contents | comments）
|
||||
|
||||
Returns:
|
||||
|
||||
"""
|
||||
|
||||
return (
|
||||
f"{self.json_store_path}/{crawler_type_var.get()}_{store_type}_{utils.get_current_date()}.json",
|
||||
f"{self.words_store_path}/{crawler_type_var.get()}_{store_type}_{utils.get_current_date()}"
|
||||
)
|
||||
|
||||
async def save_data_to_json(self, save_item: Dict, store_type: str):
|
||||
"""
|
||||
Below is a simple way to save it in json format.
|
||||
Args:
|
||||
save_item: save content dict info
|
||||
store_type: Save type contains content and comments（contents | comments）
|
||||
|
||||
Returns:
|
||||
|
||||
"""
|
||||
pathlib.Path(self.json_store_path).mkdir(parents=True, exist_ok=True)
|
||||
pathlib.Path(self.words_store_path).mkdir(parents=True, exist_ok=True)
|
||||
save_file_name,words_file_name_prefix = self.make_save_file_name(store_type=store_type)
|
||||
save_data = []
|
||||
|
||||
async with self.lock:
|
||||
if os.path.exists(save_file_name):
|
||||
async with aiofiles.open(save_file_name, 'r', encoding='utf-8') as file:
|
||||
save_data = json.loads(await file.read())
|
||||
|
||||
save_data.append(save_item)
|
||||
async with aiofiles.open(save_file_name, 'w', encoding='utf-8') as file:
|
||||
await file.write(json.dumps(save_data, ensure_ascii=False))
|
||||
|
||||
if config.ENABLE_GET_COMMENTS and config.ENABLE_GET_WORDCLOUD:
|
||||
try:
|
||||
await self.WordCloud.generate_word_frequency_and_cloud(save_data, words_file_name_prefix)
|
||||
except:
|
||||
pass
|
||||
|
||||
async def store_content(self, content_item: Dict):
|
||||
"""
|
||||
content JSON storage implementation
|
||||
Args:
|
||||
content_item:
|
||||
|
||||
Returns:
|
||||
|
||||
"""
|
||||
await self.save_data_to_json(content_item, "contents")
|
||||
|
||||
async def store_comment(self, comment_item: Dict):
|
||||
"""
|
||||
comment JSON storage implementation
|
||||
Args:
|
||||
comment_item:
|
||||
|
||||
Returns:
|
||||
|
||||
"""
|
||||
await self.save_data_to_json(comment_item, "comments")
|
||||
|
||||
async def store_creator(self, creator: Dict):
|
||||
"""
|
||||
creator JSON storage implementation
|
||||
Args:
|
||||
creator:
|
||||
|
||||
Returns:
|
||||
|
||||
"""
|
||||
await self.save_data_to_json(creator, "creators")
|
||||
|
||||
async def store_contact(self, contact_item: Dict):
|
||||
"""
|
||||
creator contact JSON storage implementation
|
||||
Args:
|
||||
contact_item: creator's contact item dict
|
||||
|
||||
Returns:
|
||||
|
||||
"""
|
||||
|
||||
await self.save_data_to_json(save_item=contact_item, store_type="contacts")
|
||||
|
||||
async def store_dynamic(self, dynamic_item: Dict):
|
||||
"""
|
||||
creator dynamic JSON storage implementation
|
||||
Args:
|
||||
dynamic_item: creator's contact item dict
|
||||
|
||||
Returns:
|
||||
|
||||
"""
|
||||
|
||||
await self.save_data_to_json(save_item=dynamic_item, store_type="dynamics")
|
||||
|
||||
|
||||
class BiliSqliteStoreImplement(AbstractStore):
|
||||
async def store_content(self, content_item: Dict):
|
||||
"""
|
||||
Bilibili content SQLite storage implementation
|
||||
Args:
|
||||
content_item: content item dict
|
||||
|
||||
Returns:
|
||||
|
||||
"""
|
||||
|
||||
from .bilibili_store_sql import (add_new_content,
|
||||
query_content_by_content_id,
|
||||
update_content_by_content_id)
|
||||
video_id = content_item.get("video_id")
|
||||
video_detail: Dict = await query_content_by_content_id(content_id=video_id)
|
||||
if not video_detail:
|
||||
content_item["add_ts"] = utils.get_current_timestamp()
|
||||
await add_new_content(content_item)
|
||||
else:
|
||||
await update_content_by_content_id(video_id, content_item=content_item)
|
||||
|
||||
async def store_comment(self, comment_item: Dict):
|
||||
"""
|
||||
Bilibili comment SQLite storage implementation
|
||||
Args:
|
||||
comment_item: comment item dict
|
||||
|
||||
Returns:
|
||||
|
||||
"""
|
||||
|
||||
from .bilibili_store_sql import (add_new_comment,
|
||||
query_comment_by_comment_id,
|
||||
update_comment_by_comment_id)
|
||||
comment_id = comment_item.get("comment_id")
|
||||
comment_detail: Dict = await query_comment_by_comment_id(comment_id=comment_id)
|
||||
if not comment_detail:
|
||||
comment_item["add_ts"] = utils.get_current_timestamp()
|
||||
await add_new_comment(comment_item)
|
||||
else:
|
||||
await update_comment_by_comment_id(comment_id, comment_item=comment_item)
|
||||
|
||||
async def store_creator(self, creator: Dict):
|
||||
"""
|
||||
Bilibili creator SQLite storage implementation
|
||||
Args:
|
||||
creator: creator item dict
|
||||
|
||||
Returns:
|
||||
|
||||
"""
|
||||
|
||||
from .bilibili_store_sql import (add_new_creator,
|
||||
query_creator_by_creator_id,
|
||||
update_creator_by_creator_id)
|
||||
creator_id = creator.get("user_id")
|
||||
creator_detail: Dict = await query_creator_by_creator_id(creator_id=creator_id)
|
||||
if not creator_detail:
|
||||
creator["add_ts"] = utils.get_current_timestamp()
|
||||
await add_new_creator(creator)
|
||||
else:
|
||||
await update_creator_by_creator_id(creator_id, creator_item=creator)
|
||||
|
||||
async def store_contact(self, contact_item: Dict):
|
||||
"""
|
||||
Bilibili contact SQLite storage implementation
|
||||
Args:
|
||||
contact_item: contact item dict
|
||||
|
||||
Returns:
|
||||
|
||||
"""
|
||||
|
||||
from .bilibili_store_sql import (add_new_contact,
|
||||
query_contact_by_up_and_fan,
|
||||
update_contact_by_id, )
|
||||
|
||||
up_id = contact_item.get("up_id")
|
||||
fan_id = contact_item.get("fan_id")
|
||||
contact_detail: Dict = await query_contact_by_up_and_fan(up_id=up_id, fan_id=fan_id)
|
||||
if not contact_detail:
|
||||
contact_item["add_ts"] = utils.get_current_timestamp()
|
||||
await add_new_contact(contact_item)
|
||||
else:
|
||||
key_id = contact_detail.get("id")
|
||||
await update_contact_by_id(id=key_id, contact_item=contact_item)
|
||||
|
||||
async def store_dynamic(self, dynamic_item):
|
||||
"""
|
||||
Bilibili dynamic SQLite storage implementation
|
||||
Args:
|
||||
dynamic_item: dynamic item dict
|
||||
|
||||
Returns:
|
||||
|
||||
"""
|
||||
|
||||
from .bilibili_store_sql import (add_new_dynamic,
|
||||
query_dynamic_by_dynamic_id,
|
||||
update_dynamic_by_dynamic_id)
|
||||
|
||||
dynamic_id = dynamic_item.get("dynamic_id")
|
||||
dynamic_detail = await query_dynamic_by_dynamic_id(dynamic_id=dynamic_id)
|
||||
if not dynamic_detail:
|
||||
dynamic_item["add_ts"] = utils.get_current_timestamp()
|
||||
await add_new_dynamic(dynamic_item)
|
||||
else:
|
||||
await update_dynamic_by_dynamic_id(dynamic_id, dynamic_item=dynamic_item)
|
||||
@@ -0,0 +1,253 @@
|
||||
# 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。
|
||||
|
||||
|
||||
# -*- coding: utf-8 -*-
|
||||
# @Author : relakkes@gmail.com
|
||||
# @Time : 2024/4/6 15:30
|
||||
# @Desc : sql接口集合
|
||||
|
||||
from typing import Dict, List, Union
|
||||
|
||||
from async_db import AsyncMysqlDB
|
||||
from async_sqlite_db import AsyncSqliteDB
|
||||
from var import media_crawler_db_var
|
||||
|
||||
|
||||
async def query_content_by_content_id(content_id: str) -> Dict:
|
||||
"""
|
||||
查询一条内容记录（xhs的帖子 ｜ 抖音的视频 ｜ 微博 ｜ 快手视频 ...）
|
||||
Args:
|
||||
content_id:
|
||||
|
||||
Returns:
|
||||
|
||||
"""
|
||||
async_db_conn: Union[AsyncMysqlDB, AsyncSqliteDB] = media_crawler_db_var.get()
|
||||
sql: str = f"select * from bilibili_video where video_id = '{content_id}'"
|
||||
rows: List[Dict] = await async_db_conn.query(sql)
|
||||
if len(rows) > 0:
|
||||
return rows[0]
|
||||
return dict()
|
||||
|
||||
|
||||
async def add_new_content(content_item: Dict) -> int:
|
||||
"""
|
||||
新增一条内容记录（xhs的帖子 ｜ 抖音的视频 ｜ 微博 ｜ 快手视频 ...）
|
||||
Args:
|
||||
content_item:
|
||||
|
||||
Returns:
|
||||
|
||||
"""
|
||||
async_db_conn: Union[AsyncMysqlDB, AsyncSqliteDB] = media_crawler_db_var.get()
|
||||
last_row_id: int = await async_db_conn.item_to_table("bilibili_video", content_item)
|
||||
return last_row_id
|
||||
|
||||
|
||||
async def update_content_by_content_id(content_id: str, content_item: Dict) -> int:
|
||||
"""
|
||||
更新一条记录（xhs的帖子 ｜ 抖音的视频 ｜ 微博 ｜ 快手视频 ...）
|
||||
Args:
|
||||
content_id:
|
||||
content_item:
|
||||
|
||||
Returns:
|
||||
|
||||
"""
|
||||
async_db_conn: Union[AsyncMysqlDB, AsyncSqliteDB] = media_crawler_db_var.get()
|
||||
effect_row: int = await async_db_conn.update_table("bilibili_video", content_item, "video_id", content_id)
|
||||
return effect_row
|
||||
|
||||
|
||||
|
||||
async def query_comment_by_comment_id(comment_id: str) -> Dict:
|
||||
"""
|
||||
查询一条评论内容
|
||||
Args:
|
||||
comment_id:
|
||||
|
||||
Returns:
|
||||
|
||||
"""
|
||||
async_db_conn: Union[AsyncMysqlDB, AsyncSqliteDB] = media_crawler_db_var.get()
|
||||
sql: str = f"select * from bilibili_video_comment where comment_id = '{comment_id}'"
|
||||
rows: List[Dict] = await async_db_conn.query(sql)
|
||||
if len(rows) > 0:
|
||||
return rows[0]
|
||||
return dict()
|
||||
|
||||
|
||||
async def add_new_comment(comment_item: Dict) -> int:
|
||||
"""
|
||||
新增一条评论记录
|
||||
Args:
|
||||
comment_item:
|
||||
|
||||
Returns:
|
||||
|
||||
"""
|
||||
async_db_conn: Union[AsyncMysqlDB, AsyncSqliteDB] = media_crawler_db_var.get()
|
||||
last_row_id: int = await async_db_conn.item_to_table("bilibili_video_comment", comment_item)
|
||||
return last_row_id
|
||||
|
||||
|
||||
async def update_comment_by_comment_id(comment_id: str, comment_item: Dict) -> int:
|
||||
"""
|
||||
更新增一条评论记录
|
||||
Args:
|
||||
comment_id:
|
||||
comment_item:
|
||||
|
||||
Returns:
|
||||
|
||||
"""
|
||||
async_db_conn: Union[AsyncMysqlDB, AsyncSqliteDB] = media_crawler_db_var.get()
|
||||
effect_row: int = await async_db_conn.update_table("bilibili_video_comment", comment_item, "comment_id", comment_id)
|
||||
return effect_row
|
||||
|
||||
|
||||
async def query_creator_by_creator_id(creator_id: str) -> Dict:
|
||||
"""
|
||||
查询up主信息
|
||||
Args:
|
||||
creator_id:
|
||||
|
||||
Returns:
|
||||
|
||||
"""
|
||||
async_db_conn: Union[AsyncMysqlDB, AsyncSqliteDB] = media_crawler_db_var.get()
|
||||
sql: str = f"select * from bilibili_up_info where user_id = '{creator_id}'"
|
||||
rows: List[Dict] = await async_db_conn.query(sql)
|
||||
if len(rows) > 0:
|
||||
return rows[0]
|
||||
return dict()
|
||||
|
||||
|
||||
async def add_new_creator(creator_item: Dict) -> int:
|
||||
"""
|
||||
新增up主信息
|
||||
Args:
|
||||
creator_item:
|
||||
|
||||
Returns:
|
||||
|
||||
"""
|
||||
async_db_conn: Union[AsyncMysqlDB, AsyncSqliteDB] = media_crawler_db_var.get()
|
||||
last_row_id: int = await async_db_conn.item_to_table("bilibili_up_info", creator_item)
|
||||
return last_row_id
|
||||
|
||||
|
||||
async def update_creator_by_creator_id(creator_id: str, creator_item: Dict) -> int:
|
||||
"""
|
||||
更新up主信息
|
||||
Args:
|
||||
creator_id:
|
||||
creator_item:
|
||||
|
||||
Returns:
|
||||
|
||||
"""
|
||||
async_db_conn: Union[AsyncMysqlDB, AsyncSqliteDB] = media_crawler_db_var.get()
|
||||
effect_row: int = await async_db_conn.update_table("bilibili_up_info", creator_item, "user_id", creator_id)
|
||||
return effect_row
|
||||
|
||||
|
||||
async def query_contact_by_up_and_fan(up_id: str, fan_id: str) -> Dict:
|
||||
"""
|
||||
查询一条关联关系
|
||||
Args:
|
||||
up_id:
|
||||
fan_id:
|
||||
|
||||
Returns:
|
||||
|
||||
"""
|
||||
async_db_conn: Union[AsyncMysqlDB, AsyncSqliteDB] = media_crawler_db_var.get()
|
||||
sql: str = f"select * from bilibili_contact_info where up_id = '{up_id}' and fan_id = '{fan_id}'"
|
||||
rows: List[Dict] = await async_db_conn.query(sql)
|
||||
if len(rows) > 0:
|
||||
return rows[0]
|
||||
return dict()
|
||||
|
||||
|
||||
async def add_new_contact(contact_item: Dict) -> int:
|
||||
"""
|
||||
新增关联关系
|
||||
Args:
|
||||
contact_item:
|
||||
|
||||
Returns:
|
||||
|
||||
"""
|
||||
async_db_conn: Union[AsyncMysqlDB, AsyncSqliteDB] = media_crawler_db_var.get()
|
||||
last_row_id: int = await async_db_conn.item_to_table("bilibili_contact_info", contact_item)
|
||||
return last_row_id
|
||||
|
||||
|
||||
async def update_contact_by_id(id: str, contact_item: Dict) -> int:
|
||||
"""
|
||||
更新关联关系
|
||||
Args:
|
||||
id:
|
||||
contact_item:
|
||||
|
||||
Returns:
|
||||
|
||||
"""
|
||||
async_db_conn: Union[AsyncMysqlDB, AsyncSqliteDB] = media_crawler_db_var.get()
|
||||
effect_row: int = await async_db_conn.update_table("bilibili_contact_info", contact_item, "id", id)
|
||||
return effect_row
|
||||
|
||||
|
||||
async def query_dynamic_by_dynamic_id(dynamic_id: str) -> Dict:
|
||||
"""
|
||||
查询一条动态信息
|
||||
Args:
|
||||
dynamic_id:
|
||||
|
||||
Returns:
|
||||
|
||||
"""
|
||||
async_db_conn: Union[AsyncMysqlDB, AsyncSqliteDB] = media_crawler_db_var.get()
|
||||
sql: str = f"select * from bilibili_up_dynamic where dynamic_id = '{dynamic_id}'"
|
||||
rows: List[Dict] = await async_db_conn.query(sql)
|
||||
if len(rows) > 0:
|
||||
return rows[0]
|
||||
return dict()
|
||||
|
||||
|
||||
async def add_new_dynamic(dynamic_item: Dict) -> int:
|
||||
"""
|
||||
新增动态信息
|
||||
Args:
|
||||
dynamic_item:
|
||||
|
||||
Returns:
|
||||
|
||||
"""
|
||||
async_db_conn: Union[AsyncMysqlDB, AsyncSqliteDB] = media_crawler_db_var.get()
|
||||
last_row_id: int = await async_db_conn.item_to_table("bilibili_up_dynamic", dynamic_item)
|
||||
return last_row_id
|
||||
|
||||
|
||||
async def update_dynamic_by_dynamic_id(dynamic_id: str, dynamic_item: Dict) -> int:
|
||||
"""
|
||||
更新动态信息
|
||||
Args:
|
||||
dynamic_id:
|
||||
dynamic_item:
|
||||
|
||||
Returns:
|
||||
|
||||
"""
|
||||
async_db_conn: Union[AsyncMysqlDB, AsyncSqliteDB] = media_crawler_db_var.get()
|
||||
effect_row: int = await async_db_conn.update_table("bilibili_up_dynamic", dynamic_item, "dynamic_id", dynamic_id)
|
||||
return effect_row
|
||||
@@ -0,0 +1,68 @@
|
||||
# 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。
|
||||
|
||||
# -*- coding: utf-8 -*-
|
||||
# @Author : helloteemo
|
||||
# @Time : 2024/7/12 20:01
|
||||
# @Desc : bilibili 媒体保存
|
||||
import pathlib
|
||||
from typing import Dict
|
||||
|
||||
import aiofiles
|
||||
|
||||
from base.base_crawler import AbstractStoreImage, AbstractStoreVideo
|
||||
from tools import utils
|
||||
|
||||
|
||||
class BilibiliVideo(AbstractStoreVideo):
|
||||
video_store_path: str = "data/bilibili/videos"
|
||||
|
||||
async def store_video(self, video_content_item: Dict):
|
||||
"""
|
||||
store content
|
||||
|
||||
Args:
|
||||
video_content_item:
|
||||
|
||||
Returns:
|
||||
|
||||
"""
|
||||
await self.save_video(video_content_item.get("aid"), video_content_item.get("video_content"), video_content_item.get("extension_file_name"))
|
||||
|
||||
def make_save_file_name(self, aid: str, extension_file_name: str) -> str:
|
||||
"""
|
||||
make save file name by store type
|
||||
|
||||
Args:
|
||||
aid: aid
|
||||
extension_file_name: video filename with extension
|
||||
|
||||
Returns:
|
||||
|
||||
"""
|
||||
return f"{self.video_store_path}/{aid}/{extension_file_name}"
|
||||
|
||||
async def save_video(self, aid: int, video_content: str, extension_file_name="mp4"):
|
||||
"""
|
||||
save video to local
|
||||
|
||||
Args:
|
||||
aid: aid
|
||||
video_content: video content
|
||||
extension_file_name: video filename with extension
|
||||
|
||||
Returns:
|
||||
|
||||
"""
|
||||
pathlib.Path(self.video_store_path + "/" + str(aid)).mkdir(parents=True, exist_ok=True)
|
||||
save_file_name = self.make_save_file_name(str(aid), extension_file_name)
|
||||
async with aiofiles.open(save_file_name, 'wb') as f:
|
||||
await f.write(video_content)
|
||||
utils.logger.info(f"[BilibiliVideoImplement.save_video] save save_video {save_file_name} success ...")
|
||||
Reference in New Issue
Block a user