Talkish is a telegram bot that help customers.

• collect feedback
• report damage order
• update location
• support Arabic and English

In order to use the bot, follow this instruction

• search in telegram about Talkish Agent
• open it
• click o start
• the bot will guide you
• if you reach the end or want to start the bot from the beginning, type /start

I used Node.js, Telegraf lib, and MongoDB inside docker.

• Node.js to run node server to connect to telegram as well as exporting HTTP enpoints to check saved data in DB
• Telegraf lib, to build the bot workflow and responses
• MongoDB as session managment and to store data, such as customer data and feedback data
• all of them shipped in Docker container
• I generated randoms order IDs for customer to pick form

get project from GitHub

cd talkish-telegram-bot

go to the following path and add the telegram group token.

/src/config/settings.ts

go back to the main folder and run Docker build

docker-compose build

run Docker up and wait until you see this message in the terminal 'DB Connected' then you can go to Bot

docker-compose up

to get all feedback records

curl -X GET localhost:4000/get_feedback

to get all customers records

curl -X GET localhost:4000/get_customers

• remove all hard coded data such as status codes, messages ..etc
• personally, I prefer to create generic class for alll common CRUD operation, and use this class instead of repating codes
• I will add Setup script to make sure that MongoDB is set properly (DB and collections created) as well as a teardown script, that will remove collections if needed
• build mechanism and middleware for Error Handling
• implment security mechanism for the whole system.
• I will add BOT_TOEKN to .gitignore and feed this info from process.env
• add inline query to the bot, fetching helpful information from FAQ
• image path has the bot token, for the real scenario, this is a risk! so, I would upload images in private storage and use this path instead of relying on telegram storage.
• I will use i18n lib for supporting multi-langauages

• I used telegraf-session-mongodb lib for session managment, which turn out lacked option to implment expire value for session /index collection so if I had more time, I would write my own telegraf sessions managment MongoDB middleware.

• I used JS models files because TS files throw error in the librray that I used and due to time constraint, I decided to use JS