开始动手前先停下想想——这些历史信息就像旧照片，也是宝贵的财产，为什么非要删了呢？
为了消除令人尴尬的黑历史，还是为了保护个人隐私？
如果你单纯是不想要号了，可以直接去注销账号，那样更快。
也许你只是想进行一次数字断舍离，和过去的自己做个决裂……好吧，这也不是坏事。
不过删除就无法恢复了，一定请想好之后再动手，本人曾经就把QQ相册清空导致至今后悔不已。三思！
Web技术原理我就不说了，我不是专家也说不清楚，既然你能看到这你肯定是有基础的。
按我自己的理解，这种自动化脚本有两种思路：
或者更普通的思路，1和2的结合，也是本文的思路。
本脚本的运行过程是这样：
存在的问题：
删除脚本：
import asyncio
from playwright.async_api import async_playwright
import random
from auto_scripts.bvid_aid_conversion import av2bv
async def get_replied_comments(context, max_num=20):
async def get_single(id=None, reply_time=None):
params = [
"platform=web",
"build=0",
"mobi_app=web",
"web_location=333.40164"
]
if id and reply_time:
params.append(f"id={id}")
params.append(f"reply_time={reply_time}")
request_url = f"https://api.bilibili.com/x/msgfeed/reply?{'&'.join(params)}"
try:
res_json = await context.pages[0].evaluate(f"""
async () => {{
const res = await fetch('{request_url}', {{
method: 'GET',
headers: {{
'Referer': 'https://message.bilibili.com/',
'Origin': 'https://message.bilibili.com'
}},
credentials: 'include'
}});
return await res.json();
}}
""")
res_json_data = res_json.get('data')
items = res_json_data.get("items")
is_end = res_json_data.get("cursor").get("is_end")
next_id = res_json_data.get("cursor").get("id")
next_reply_time = res_json_data.get("cursor").get("time")
return items, is_end, next_id, next_reply_time
except Exception as e:
print(f"获取“回复我的”失败: {e}")
return [], True, None, None
items, is_end, next_id, next_reply_time = await get_single()
while not is_end:
if len(items) >= max_num:
break
item, is_end, next_id, next_reply_time = await get_single(next_id, next_reply_time)
if len(item) > 0:
items.extend(item)
print(f"已获取回复评论条目: {len(items)}")
await asyncio.sleep(1)
return items
async def get_liked_comments(context, max_num=20):
async def get_single(id=None, like_time=None):
params = [
"platform=web",
"build=0",
"mobi_app=web",
"web_location=333.40164"
]
if id and like_time:
params.append(f"id={id}")
params.append(f"like_time={like_time}")
request_url = f"https://api.bilibili.com/x/msgfeed/like?{'&'.join(params)}"
try:
res_json = await context.pages[0].evaluate(f"""
async () => {{
const res = await fetch('{request_url}', {{
method: 'GET',
headers: {{
'Referer': 'https://message.bilibili.com/',
'Origin': 'https://message.bilibili.com'
}},
credentials: 'include'
}});
return await res.json();
}}
""")
res_json_data = res_json.get('data').get("total")
items = res_json_data.get("items")
is_end = res_json_data.get("cursor").get("is_end")
next_id = res_json_data.get("cursor").get("id")
next_like_time = res_json_data.get("cursor").get("time")
return items, is_end, next_id, next_like_time
except Exception as e:
print(f"获取“回复我的”失败: {e}")
return [], True, None, None
items, is_end, next_id, next_like_time = await get_single()
while not is_end:
if len(items) >= max_num:
break
item, is_end, next_id, next_like_time = await get_single(next_id, next_like_time)
if len(item) > 0:
items.extend(item)
print(f"已获取被赞评论条目: {len(items)}")
await asyncio.sleep(1)
return items
async def get_csrf(context):
cookies = await context.cookies("https://www.bilibili.com")
bili_jct_cookie = next((c for c in cookies if c['name'] == 'bili_jct'), None)
return bili_jct_cookie['value'] if bili_jct_cookie else None
async def delete_notification(context, notification_id, tp):
# tp=0 被赞通知， tp=1 被回复通知
csrf_value = await get_csrf(context)
try:
response = await context.pages[0].evaluate(f"""
fetch('https://api.bilibili.com/x/msgfeed/del', {{
method: 'POST',
headers: {{
'Content-Type': 'application/x-www-form-urlencoded',
'Referer': 'https://message.bilibili.com/',
'Origin': 'https://message.bilibili.com',
'Sec-Fetch-Mode': 'cors',
'Sec-Fetch-Site': 'same-origin'
}},
body: new URLSearchParams({{
'tp': '{tp}',
'id': '{notification_id}',
'build': '0',
'mobi_app': 'web',
'csrf': '{csrf_value}'
}}),
credentials: 'include'
}}).then(res => res.json())
""")
return response.get('code') == 0
except Exception as e:
print(f"删除“回复我的”通知失败: {e}")
return False
async def delete_comment(context, aid, comment_id, vd_source):
referer = f"https://www.bilibili.com/video/{av2bv(aid)}/?vd_source={vd_source}"
csrf_value = await get_csrf(context)
try:
response = await context.pages[0].evaluate(f"""
fetch('https://api.bilibili.com/x/v2/reply/del', {{
method: 'POST',
headers: {{
'Content-Type': 'application/x-www-form-urlencoded',
'Referer': '{referer}',
'Origin': 'https://www.bilibili.com',
'Sec-Fetch-Mode': 'cors',
'Sec-Fetch-Site': 'same-origin'
}},
body: new URLSearchParams({{
'oid': '{aid}',
'type': '1',
'rpid': '{comment_id}',
'csrf': '{csrf_value}'
}}),
credentials: 'include'
}}).then(res => res.json())
""")
return response.get('code') == 0
except Exception as e:
print(f"删除“回复我的”评论失败: {e}")
return False
async def delete_danmu(context, aid, dmid, cid, vd_source, dm_progress):
referer = f"https://www.bilibili.com/video/{av2bv(aid)}/?dm_progress={dm_progress}&dmid={dmid}&vd_source={vd_source}"
csrf_value = await get_csrf(context)
try:
response = await context.pages[0].evaluate(f"""
fetch('https://api.bilibili.com/x/v2/dm/recall', {{
method: 'POST',
headers: {{
'Content-Type': 'application/x-www-form-urlencoded',
'Referer': '{referer}',
'Origin': 'https://www.bilibili.com',
'Sec-Fetch-Mode': 'cors',
'Sec-Fetch-Site': 'same-origin'
}},
body: new URLSearchParams({{
'dmid': '{dmid}',
'cid': '{cid}',
'type': '1',
'csrf': '{csrf_value}'
}}),
credentials: 'include'
}}).then(res => res.json())
""")
return response.get('code') == 0
except Exception as e:
print(f"删除弹幕失败: {e}")
return False
async def delete_replied_comments(context, items, vd_source):
processed_comment = set()
for item in items:
if item.get("item").get("business") == "评论":
comment_id = item.get("item").get("target_id")
my_comment = item.get("item").get("target_reply_content")[:20]
if my_comment == "":
my_comment = item.get("item").get("root_reply_content")[:20]
print(f"\n---\n评论内容：{my_comment}... || id: '{comment_id}'")
# input("请按回车键确认删除...")
if comment_id not in processed_comment:
comment_deleted = await delete_comment(context, item.get("item").get("subject_id"), comment_id, vd_source)
print(f"评论: '{comment_id}' 删除状态: {comment_deleted}")
if not comment_deleted:
print(f"删除失败：{item}")
processed_comment.add(comment_id)
else:
print("评论已经处理过，跳过")
else:
print(f"预期外的类型，跳过: {item.get("item").get("business")}")
# 删除对应的通知
ok = await delete_notification(context, item.get("id"), 1)
print(f"通知: '{item.get("id")}' 删除状态: {ok}")
await asyncio.sleep(random.randint(0, 1))
async def delete_liked_commets(context, items, vd_source):
processed_comment = set()
for item in items:
if item.get("item").get("business") == "评论":
comment_id = item.get("item").get("item_id")
my_comment = item.get("item").get("title")[:20]
print(f"\n---\n评论内容：{my_comment}... || id: '{comment_id}'")
# input("请按回车键确认删除...")
# 从native_uri中提取视频aid，uri形如：bilibili://video/{xxxx}?...
native_uri = item.get("item").get("native_uri")
import re
pattern = r"bilibili://video/(\d+)" # 仅处理了视频评论，如果是动态评论可以改下pattern，此时native_uri例如'bilibili://comment/detail/11/381146508/288325939232/?subType=0&anchor=288326706768&showEnter=1&extraIntentId=0&scene=1&enterName=查看动态详情&enterUri=bilibili://opus/detail/1156902561345699849'
match = re.search(pattern, native_uri)
if not match:
print(f"删除失败：未能提取oid，native_uri={native_uri}")
continue
aid = int(match.group(1))
# 开始删除评论
if comment_id not in processed_comment:
comment_deleted = await delete_comment(context, aid, comment_id, vd_source)
print(f"评论: '{comment_id}' 删除状态: {comment_deleted}")
if not comment_deleted:
print(f"删除失败：{item}")
processed_comment.add(comment_id)
else:
print("评论已经处理过，跳过")
elif item.get("item").get("business") == "弹幕":
# 从native_uri中提取 aid, dm_progress, cid, dmid
native_uri = item.get("item").get("native_uri")
import re
pattern = r"bilibili://video/(\d+)\?dm_progress=(\d+)&cid=(\d+)&dmid=(\d+)"
match = re.search(pattern, native_uri)
if not match:
print(f"删除失败：未能提取oid，native_uri={native_uri}")
continue
aid = int(match.group(1))
dm_progress = int(match.group(2))
cid = int(match.group(3))
dmid = int(match.group(4))
print(f"\n---\n弹幕内容：{item.get("item").get("title")}... || id: '{dmid}'")
# 开始删除弹幕
if dmid not in processed_comment:
danmu_deleted = await delete_danmu(context, aid, dmid, cid, vd_source, dm_progress)
print(f"弹幕: '{dmid}' 删除状态: {danmu_deleted}")
if not danmu_deleted:
print(f"删除失败：{item}")
processed_comment.add(dmid)
else:
print("弹幕已经处理过，跳过")
# 删除对应的通知
ok = await delete_notification(context, item.get("id"), 0)
print(f"通知: '{item.get("id")}' 删除状态: {ok}")
await asyncio.sleep(random.randint(0, 1))
async def run_cleaner():
async with async_playwright() as p:
user_data_dir = "./bilibili_user_data"
context = await p.chromium.launch_persistent_context(
user_data_dir,
headless=False, # 必须为 False，我们要手动登录
args=["--start-maximized"],
no_viewport=True
)
page = context.pages[0]
# 获取vd_source
await page.goto("https://www.bilibili.com/video/BV1NQtDeEErj") # 随便打开一个视频
await page.wait_for_function(
"window.location.href.includes('vd_source')",
timeout=10000
)
vd_source = page.url.split("vd_source=")[1]
await asyncio.sleep(1)
await page.goto("https://message.bilibili.com")
await page.wait_for_load_state(state="domcontentloaded")
# 被回复的评论
replied_items = await get_replied_comments(context, 10000)
await delete_replied_comments(context, replied_items, vd_source)
# 被点赞的评论和弹幕
for i in range(10):
liked_items = await get_liked_comments(context, 200)
await delete_liked_commets(context, liked_items, vd_source)
# 任务完成后保持一会儿
await asyncio.sleep(1)
await context.close()
if __name__ == "__main__":
asyncio.run(run_cleaner())
BV号 <=> aid 转换，可以调接口，但有人已经把计算代码公开了：
XOR_CODE = 23442827791579
MASK_CODE = 2251799813685247
MAX_AID = 1 << 51
ALPHABET = "FcwAPNKTMug3GV5Lj7EJnHpWsx4tb8haYeviqBz6rkCy12mUSDQX9RdoZf"
ENCODE_MAP = 8, 7, 0, 5, 1, 3, 2, 4, 6
DECODE_MAP = tuple(reversed(ENCODE_MAP))
BASE = len(ALPHABET)
PREFIX = "BV1"
PREFIX_LEN = len(PREFIX)
CODE_LEN = len(ENCODE_MAP)
def av2bv(aid: int) -> str:
bvid = [""] * 9
tmp = (MAX_AID | aid) ^ XOR_CODE
for i in range(CODE_LEN):
bvid[ENCODE_MAP[i]] = ALPHABET[tmp % BASE]
tmp //= BASE
return PREFIX + "".join(bvid)
def bv2av(bvid: str) -> int:
assert bvid[:3] == PREFIX
bvid = bvid[3:]
tmp = 0
for i in range(CODE_LEN):
idx = ALPHABET.index(bvid[DECODE_MAP[i]])
tmp = tmp * BASE + idx
return (tmp & MASK_CODE) ^ XOR_CODE
assert av2bv(113179954776959) == "BV1NQtDeEErj"
assert bv2av("BV1NQtDeEErj") == 113179954776959
好，你应该已经删完了吧？我现在有一个好（坏？）消息要告诉你。
好消息是，如果你以后后悔了，想找回被自己手贱删除的历史，可以在这个 aicu.cc 网站查到。
这是一个非官方站点，爬取了整个B站所有的公开评论/弹幕（B站官方你们的技术团队在干什么……不过想想上次整个站点源代码被上传github这种事，站点被爬又算得了什么）。
虽然不能恢复，但它也提供了一种历史查看手段。
但这同时也是一个坏消息，如果你删历史是为了隐藏自己的黑历史，那你失算了，任何人都可以轻易在这个站点上查到你的黑历史。享受裸奔的快乐吧ヾ(o◕∀◕)ﾉヾ。