This is an early release version, everything are subject to change, please DO NOT used for production!!!
这是一个提前释出的版本,所有内容都可能会改变,请不要用于生产环境
This repository included core/crawler/api/scripts
, frontend repository is here.
这个仓库包含 核心/爬虫/api/脚本
,前端仓库位于这里
- x86_64/arm64
- Node.js 18+
- Nginx
- MySQL 8.0 / MariaDB (some function is not supported MariaDB, like function
ANY_VALUE()
, parserngram
) / SQLite
We use GitHub Actions to build some scripts, you can find them in BANKA2017/twitter-monitor/-/Actions/build_rollup (Login required)
- leave it as is if not necessary
- service tmv1 and analytics in
SQL_CONFIG
are not necessary, execpt you used twitter monitor before 2020-03
- NO TYPESCRIPT, core code are not yet supported typescript
- copy and rename the setting file from
libs/assets/settings_sample.mjs
tolibs/assets/settings.mjs
, and edit it - enjoy it!
IF POSSIBLE, DO NOT EXECUTE THOSE SCRIPTS
- execute
node apps/scripts/updateQueryIdList.mjs
to updatequeryId
- execute
node apps/scripts/checkGraphqlFeaturesStatus.mjs
for check after updatingqueryId
-
edit
SQL_CONFIG
to enable service twitter_monitor -
execute
yarn run init
ornpm run init
-
open
config.html
by browser to edit and save config fileconfig.json
aslibs/assets/config.json
-
PM2 is a good choice for you, you also can use screen or nohup
pm2 start apps/crawler/get_tweets.mjs
-
set crontab,e.g.
#those are example path, use your path * * * * * node apps/crawler/blurhash.mjs #PHP version is better, node version is slow */10 * * * * node apps/crawler/updatePollsAndAudioSpace.mjs
-
you can set
TWEETS_SAVE_PATH
to save tweet content as json in that path -
we use another bearer authorization token to crawl more tweets, but this token is not supported some new feature like mix media, you can replace it by the latest token (then not support nsfw content) -
Account pools are not yet supported. In the future, you will need to prepare a large number of accounts to build account pools.
- execute
yarn run dbapi
to enable the api only used databases (tmv1, twitter monitor) - execute
yarn run api
to enable full version api (dbapi and album api, online api, media proxy) - You need to prepare a large number of accounts to build an account pool. To learn more about account pool, please read open_account/README.md. Then copy the
guest_accounts.json
created by the script to the project root directory/the same directory as theapp.mjs
/app_online.mjs
, or paste its content into the variableGUEST_ACCOUNTS
oflibs/assets/setting.mjs
- *you can create a better api than mine
- a script in
apps/scripts
namedwatchDog.mjs
can be added to crontab to check whether some data being added in latest 30 minites
- archive userinfo, most tweets(included reply) and nearly ALL MEDIA(included avatar and banner) by search api / timeline api,
Following
andFollowers
- spaces, boradcast (ffmpeg command)
- PLEASE PRECHECK THE ACCOUNT HAVEN BEEN SEARCHBAN
- Read more in archiver/README.md
- supported album api, online api, media proxy and translator api
- You need to prepare a large number of accounts to build an account pool. To learn more about account pool, please read open_account/README.md. Then save the content as
guest_accounts
to kv - to deploy it is easy
- *read https://developers.cloudflare.com/workers/wrangler/workers-kv/#create-a-kv-namespace-with-wrangler
- create kv space named 'twitter-monitor-workers-kv' in https://dash.cloudflare.com/ or executed
npx wrangler kv:namespace create kv
- copy the value of 'id' into 'kv_namespaces[0].id'
- execute
npx wrangler publish
-
set env
android_guest_account
likes{ "authorization": "Bearer AAAAAAAAAAAAAAAAAAAAAFXzAwAAAAAAMHCxpeSDG1gLNLghVe8d74hl6k4%3DRUMF4xAQLsbeBhTSRrCiQpJtxoGWeyHrDb5te2jpGskWDFW82F", "oauth_token": "...", "oauth_token_secret": "..." }
-
then execute
node apps/rate_limit_checker/run.mjs
-
Read more in rate-limit-checker/README.md
- Read more in open_account/README.md
- Read more in web_push/README.md
* summary*
* summary_large_image
* promo_image_convo
* promo_video_convo
* promo_website
* audio*
* player
* periscope_broadcast
* broadcast
* promo_video_website
* promo_image_app
* app
* direct_store_link_app
* live_event
* moment**
* poll2choice_text_only
* poll3choice_text_only
* poll4choice_text_only
* poll2choice_image
* poll3choice_image
* poll4choice_image
* appplayer
* audiospace
* unified_card
* image_website
* video_website
* image_carousel_website
* video_carousel_website
* image_app
* video_app
* image_carousel_app
* video_carousel_app
* image_multi_dest_carousel_website
* video_multi_dest_carousel_website
* mixed_media_multi_dest_carousel_website
* mixed_media_single_dest_carousel_website
* mixed_media_single_dest_carousel_app
* image_collection_website
* twitter_list_details
* media_with_details_horizontal (for topic ?)
* twitter_article
* community_details
AxiosHelper
used to fix the issue that CloudFlare Workers unable to use axios, Read more
GoogleTokenGenerator.php
from google-translate-php, I remove some unnecessary code //now we need't use tk in Google translate- function
get_mime
is rewrited from Tieba-Cloud-Sign - you may notice a package named
sequelize-automate-gm
inpackage.json
, this is a tiny tool to help me create model from exist tables
- x86_64/arm64 (理论上是全平台支持的)
- Node.js 18+
- Nginx
- MySQL 8.0 (使用MariaDB等基于 MySQL 5.x 的发行版可能不能支持函数
ANY_VALUE()
, 解析器ngram
,会导致部分api不可用,若不使用api可忽略此处) / SQLite
- 如无必要,保持原样
SQL_CONFIG
中的 tmv1 以及 analytics 的服务都不是必须的,除非您在 2022-03 重写前就使用了 twitter monitor
- 没有 Typescript,其实是不知怎么入手
- 拷贝编辑
libs/assets/settings_sample.mjs
并另存为libs/assets/settings.mjs
- 开始使用吧!
如无必要,请不要跑这些脚本
- 运行
node apps/scripts/updateQueryIdList.mjs
以更新queryId
- 更新
queryId
后运行node apps/scripts/checkGraphqlFeaturesStatus.mjs
进行检查(不通过怎么办?可以来提issue)
-
编辑
SQL_CONFIG
以启用本服务 -
运行
yarn run init
或npm run init
以导入数据表以及在路径libs/assets/config.json
创建配置文件 -
我推荐使用 PM2,但 nohup 和 screen 也不是不行
pm2 start apps/crawler/get_tweets.mjs
-
编辑crontab
#这些都是示例,用你自己的路径 * * * * * node apps/crawler/blurhash.mjs #PHP版更好 */10 * * * * node apps/crawler/updatePollsAndAudioSpace.mjs
-
设定
TWEETS_SAVE_PATH
可以将推文内容保存为json文件 -
我们使用旧版bearer authorization token以获得更多的推文,但这个token并不支持一些新版的特性,比如混合媒体,你可以用新的token替换之(然后不支持 NSFW 内容) -
目前还不支持帐号池,未来需要自行准备大量帐号构建帐号池
- 运行
yarn run dbapi
启用仅使用数据库的api (tmv1,twitter monitor) - 运行
yarn run api
启用全部 api (上一项的全部以及 album api,online api,media proxy) - 需要自行准备大量帐号构建帐号池,关于帐号池请查看 open_account/README.md。然后将脚本创建的
guest_accounts.json
拷贝到 项目根目录/执行脚本相同目录,或者将其内容粘贴到libs/assets/setting.mjs
的变量GUEST_ACCOUNTS
中 - *欢迎自行开发一个更好的api
apps/scripts
里面有一个叫watchDog.mjs
的脚本,可以添加到 crontab 中用以检查半小时内是否有数据增加
- 通过搜索API或时间线API备份帐号的用户信息,大多数推文(包括回复)和媒体文件(包括当前头像和banner),
Following
和Followers
- 备份Spaces、播客(生成 ffmpeg 命令)
- 使用前请检查待备份帐号是否被搜索封禁
- 使用方式请阅读 archiver/README.md
- 支持 album api, online api, media proxy 以及 翻译 api
- 需要自行准备大量帐号构建帐号池,关于帐号池请查看 open_account/README.md。然后将脚本创建的
guest_accounts.json
的内容命名为guest_accounts
,保存到kv
- 部署只需要简单的几步
- *(可跳过)阅读 https://developers.cloudflare.com/workers/wrangler/workers-kv/#create-a-kv-namespace-with-wrangler
- 在 https://dash.cloudflare.com/ 创建一个叫 'twitter-monitor-workers-kv' 的 kv 命名空间,或者直接执行
npx wrangler kv:namespace create kv
- 将返回的 'id' 的值拷贝到 'kv_namespaces[0].id'
- 执行
npx wrangler publish
-
设置环境变量
android_guest_account
{ "authorization": "Bearer AAAAAAAAAAAAAAAAAAAAAFXzAwAAAAAAMHCxpeSDG1gLNLghVe8d74hl6k4%3DRUMF4xAQLsbeBhTSRrCiQpJtxoGWeyHrDb5te2jpGskWDFW82F", "oauth_token": "...", "oauth_token_secret": "..." }
-
然后执行
node apps/rate_limit_checker/run.mjs
- 了解更多 web_push/README.md
-
支持的卡片类型 (
*
表示使用小图,**
表示使用小图且图片在卡片右侧)- summary*
- summary_large_image
- promo_image_convo
- promo_video_convo
- promo_website
- audio*
- player
- periscope_broadcast
- broadcast
- promo_video_website
- promo_image_app
- app
- direct_store_link_app
- live_event
- moment**
- poll2choice_text_only
- poll3choice_text_only
- poll4choice_text_only
- poll2choice_image
- poll3choice_image
- poll4choice_image
- appplayer
- audiospace
- unified_card
- image_website
- video_website
- image_carousel_website
- video_carousel_website
- image_app
- video_app
- image_carousel_app
- video_carousel_app
- image_multi_dest_carousel_website
- video_multi_dest_carousel_website
- mixed_media_multi_dest_carousel_website
- mixed_media_single_dest_carousel_website
- mixed_media_single_dest_carousel_app
- image_collection_website
- twitter_list_details
- media_with_details_horizontal (for topic ?)
- twitter_article
- community_details
AxiosHelper
用于处理解决 Workers 无法使用 axios 的问题, 阅读更多
GoogleTokenGenerator.php
这个文件出自 google-translate-php 并经过本人小幅修改; 现在使用 Google translate 不再需要生成 TKK,故已移除本文件- 函数
get_mime
重写自 Tieba-Cloud-Sign - 在
package.json
里面有一个叫sequelize-automate-gm
的包,这是我用来从现有的数据库中导出 model 用的小工具(虽然有一些奇奇怪怪的 bug)
The PHP version will be updated soon... maybe?