gasolin / saihubot Goto Github PK

to bypass the CORS issue, Make a CORS (server-side) companion
https://www.npmjs.com/package/cors-anywhere
https://www.npmjs.com/package/qorsproxy

currently can use https://cors-anywhere.herokuapp.com/ to fetch data

add option to autosave brain (by setInterval)

There's a helper plugin which currently show names of plugins, let's add help attr in each plugin and show it when the help attr is available

ex refer https://tailwindcomponents.com/component/chat

correspondent to html-adapter/saihubot-html-skill-help

port from https://github.com/mihai/hubot-currency/blob/master/src/currency.coffee via http://coffeescript.org/#try:

current structure

https://github.com/gasolin/saihubot in gh-pages branch
- index.html
- docs
- samples
- src
  - adapters
  - addons
  - core
  - skills

https://github.com/gasolin/saihubot-cli in main branch
- src
  - adapters
  - addons
  - cli.js

suggest structure: (monorepo manage by lerna)

keep commit in gh-pages branch so current html version will keep working

Name modules as saihubot-[adapter]-[addon|skill]-[name]

https://github.com/gasolin/saihubot in gh-pages branch
- index.html
- cli/ // previous https://github.com/gasolin/saihubot-cli 
- docs
- samples
- packages
  - core
  -  html-adapter
  - html-addon-search
  ...
  - html-skill-search
  ...
  - cli-adaper
  - cli-addon-search
  ...
  - cli-skill-search
  ...

include keyword saihubot, saihubot-adapter, saihubot-skill, saihubot-addon in package.json

readonly config brain, accept path

requirement:{
  brain: {
    props: ['read', 'config'],
  }
}

refer to gasolin/webbybot#67

wrap fetch api as Saihu.http

http(URL)
  .query({})
  .header('User-Agent', 'Mozilla/5.0')
  .get(err, res, body) {...}

API:

huohubot.prompt(title, callback, [fail_message])

wait input for 30 seconds

then create index.html with full features

use https://www.npmjs.com/package/os-locale to detect locale,
can take extra SAIHUBOT_LANG to overwrite the default

define in skill

i18n: {
  'en': {
    ticker: 'High {{high}} / Low {{low}}'
  },
  'zh_TW': {
     ticker: '高 {{high}} / 低 {{low}}'
  }
}

call robot.addon.t('ticker') to render string

call robot.addon.i18n.language to get current lang

ref https://www.i18next.com/translation-function/interpolation

create saihubot-brain-localstorage.js addon to map with hubot's brain module

https://github.com/mnpk/hubot-google/blob/master/src/google.coffee

https://github.com/github/hubot-scripts/blob/master/src/scripts/google.coffee

ADV:

1. as the proxy to APIs that not accept CORS
2. JS like google app script are used
3. hide sensitive token to frontend
4. able to set a scheduler to run craw script periotically

use sort of redux pattern (dispatch->reducer) and host state to reflect any change.

add rules manager to handle rules/skills

• validate
• add

Then we can implement more functions in followup bugs

• remove by id
• run by id
• gen help

https://developer.mozilla.org/zh-CN/docs/Web/JavaScript/Guide/Modules

save dialog actions in brain and process skills in brain before handle plugin skills

ex: specify webextension sync driver to localforage brain
https://github.com/esphen/localforage-webExtensionStorage-driver

get system env and return a proper type

all system env are prefixed with SAIHUBOT_

SAIHUBOT_ETH_ADDRS=addr1, addr2

run getConfig('LANG')

should return as 'en' (string)

run getConfig('ETH_ADDRS')

should return as [addr1, addr2] (array)

provide redux-adapter and reducerSlice that fit React (maybe React Native) usage

current huohubot-dialog only take care of click input, but does not parse text input

expect:

when user type something, the bot will parse it and treat as its selection

https://pouchdb.com/adapters.html
https://github.com/Level/level

wrap window.fetch / node_fetch as addons.fetch API in saihubot-[adapter]-addons-search

https://github.com/VerbalExpressions/JSVerbalExpressions

wrap robot.emit(CustomEvent) and robot.on(addEventListner) method for event handling

https://github.com/hubot-scripts/hubot-google-translate/blob/master/src/google-translate.coffee

and define actions to access the brain

pick a magic number to just keep last 8 dialogues

cron like scheduler to notify by time/date

construct messages with markdown syntax

with https://github.com/vadimdemedes/ink