Hrenobot

Description

This is a Telegram bot written on Node.JS with microservices. It waits for user messages and somehow reacts to them.

Microservices

Execution flow:

• Bot receive message from Telegram chat and send it to Respondent
• Respondent extract part of the message and send it to Librarian (other microservices can be connected here)
• Librarian search in DB for reaction and send it back to Respondent
• Respondend receives answer from Librarian and send it back to Bot (may receive answers from different microservices and decide what to choose)
• if Bot received some non-null answer after all this chain then it post it in the Telegram chat.

Bot

Telegram bot itself. Uses node-telegram-bot-api.

For any new message in chat:

• Get text from message
• POST / { text: _chat_message_text_ } to Respondent microservice
• Wait for the response
• If response contains non-null answer field then take it and print in Telegram chat

Requirements

Must have .env file containing

TELEGRAM_BOT_TOKEN=telegram_token_from_bot_father

Bot commands

• /update - sends GET /update to Respondent microservice, print in Telegram chat number of updated records returned by Respondent

Respondent

RestAPI service. Receives full chat message from Bot. Decides from were and how to get answer to it. Returns { answer: _answer_text_ }. Returns { answer: null } if cannot find answer.

API

• POST /
  • Receives data in JSON { text: _text_ }
  • Get last word from _text_
  • Send it with GET /get to Librarian
  • Waits for the response, extract reaction and returns it upper as answer
• GET /update
  • Sends GET /update to the Librarian
  • Waits for the response and returns it upper

Librarian

RestAPI service. Returns reaction for passed word. Uses Google Spreadsheets as storage for word -> reaction pairs.

API

• GET /get?text=_text_
  • Looks for _text_ key in dictionary
  • Returns { reaction: _reaction_ } if found, { reaction: null } if not.
• GET /update
  • Updates internal dictionary with data from Google Spreadheet, returns number of stored records in form { records: _records_count_ }. This request takes about a second and also limited by Google API (100 requests per minute or something).

Requirements

Must have .env file containing

SPREADSHEET_ID=_google_spreadsheet_uid_
API_KEY=_api_key_for_connected_spreadsheet_api_from_google_console_

First sheet of Google Spreadsheet must have two columns named text and reaction.