引子和规划
周末的时候反思了下自己，感觉日常总是好奇新知识，而没有把看过的老知识彻底学会。所以决定停止接受外部知识一段时间，整理下以前看过，但是还一知半解的文章。想来想去，感觉需要把所有收藏的文章整合一下，然后做一个本地的知识搜索系统。大概有以下几个来源：
- 知乎
- 赞过的答案 done
- 回答
- 收藏
- 想法
- 微信收藏
- 收藏的链接
- Pocket
- 收藏的链接
- 浏览器
- 书签
- 浏览记录
- StackOverflow
- 赞过的问题
- 赞过的回答
- 收藏的问题
- GitHub
- Star 的库
- 参与过的讨论所以，第一步是先把知乎赞过的回答和文章都导出到本地。后续其他的站点再慢慢写。
查找知乎的 API
不得不说，知乎的 API 设计还是挺好的，贴合 RESTful 但又不强行 RESTful。总体也非常规整，没有什么奇葩的地方。知乎上没有单独的赞过的回答这个 API，而是在个人的 timeline 中。打开个人主页，比如我的：zhihu.com/people/kongyi。然后 Cmd+Opt+I 审查元素，随便往下滚动一点，找到请求：
知乎 API
可以看到 API 是：
https://www.zhihu.com/api/v3/moments/kongyifei/activities?limit=7&sdesktop=true
把这个请求通过浏览器的 copy as curl 右键菜单复制出来，然后到 curl2py 这个网站直接转换成 Python 请求。
编写脚本
最后，我们把刚刚那个请求简单扩充一下，编写一个脚本遍历：
"""
This script download all my activities from zhihu. The main reason is to see what I have upvoted is useful
"""
import json
import sqlite3
import time
from datetime import datetime
import requests
headers = {
"x-api-version": "3.0.40",
"user-agent": "Mozilla/5.0 (Macintosh; Intel Mac OS X 11_2_0) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/89.0.4389.90 Safari/537.36",
"x-requested-with": "fetch",
"sec-ch-ua": '"Google Chrome";v="89", "Chromium";v="89", ";Not A Brand";v="99"',
"accept": "*/*",
"referer": "https://www.zhihu.com/people/kongyifei",
"accept-language": "en,zh-CN;q=0.9,zh-TW;q=0.8,zh;q=0.7",
"cookie": '这里需要自己的 Cookie',
}
ended = False
url = "https://www.zhihu.com/api/v3/moments/kongyifei/activities?limit=7&desktop=true"
db = sqlite3.connect("zhihu.db")
c = db.cursor()
c.execute(
"""create table if not exists upvoted_answers (
id integer primary key,
time_upvoted datetime,
author text,
author_url text,
comment_count integer,
voteup_count integer,
question text,
answer text,
url text,
topic_ids text,
time_created datetime,
time_updated datetime
);
"""
)
c.execute(
"""create table if not exists upvoted_articles (
id integer primary key,
time_upvoted datetime,
author text,
author_url text,
comment_count integer,
voteup_count integer,
title text,
content text,
url text,
image_url text,
time_created datetime,
time_updated datetime
);
"""
)
while not ended:
print(url)
try:
response = requests.get(url, headers=headers,)
data = response.json()
except Exception:
print("connection blocked, wait for a few seconds...")
time.sleep(5)
continue
ended = data["paging"]["is_end"]
url = data["paging"].get("next", "")
time.sleep(0.5)
for item in data["data"]:
if item["action_text"] not in ["赞同了回答", "赞同了文章"]:
continue
if item["action_text"] == "赞同了回答":
upvote = dict(
time_upvoted=item["created_time"],
author=item["target"]["author"].get("name"),
author_url="https://zhihu.com/people/" + item["target"]["author"].get("id", ""),
comment_count=item["target"]["comment_count"],
voteup_count=item["target"]["voteup_count"],
question=item["target"]["question"]["title"],
answer=item["target"]["content"],
url="https://zhihu.com/question/%s/answer/%s"
% (item["target"]["question"]["id"], item["target"]["id"]),
topic_ids=json.dumps(item["target"]["question"]["bound_topic_ids"]),
time_created=item["target"]["created_time"],
time_updated=item["target"]["updated_time"]
)
c.execute(
"insert into upvoted_answers"
"(time_upvoted, author, author_url, comment_count, question, "
"answer, url, voteup_count, topic_ids, time_created, time_updated)"
"values"
"(:time_upvoted, :author, :author_url, :comment_count, :question, "
":answer, :url, :voteup_count, :topic_ids, :time_created, :time_updated)",
upvote,
)
print(
datetime.fromtimestamp(upvote["time_upvoted"]).strftime("%Y-%m-%d"),
upvote["question"],
)
elif item["action_text"] == "赞同了文章":
upvote = dict(
time_upvoted=item["created_time"],
author=item["target"]["author"].get("name"),
author_url="https://zhihu.com/people/" + item["target"]["author"].get("id", ""),
comment_count=item["target"]["comment_count"],
voteup_count=item["target"]["voteup_count"],
title=item["target"]["title"],
content=item["target"]["content"],
url=item["target"]["url"],
image_url=item["target"]["image_url"],
time_created=item["target"]["created"],
time_updated=item["target"]["updated"]
)
c.execute(
"insert into upvoted_articles"
"(time_upvoted, author, author_url, comment_count, title, content, "
"url, voteup_count, image_url, time_created, time_updated)"
"values"
"(:time_upvoted, :author, :author_url, :comment_count, :title, :content, "
":url, :voteup_count, :image_url, :time_created, :time_updated)",
upvote,
)
print(
datetime.fromtimestamp(upvote["time_upvoted"]).strftime("%Y-%m-%d"),
upvote["title"],
)
db.commit()
print("All set!")
运行截图
大概跑上几十分钟，我们赞过的回答和文章就备份好啦~
导出的回答
下一篇：备份自己在知乎上的所有回答。