Converse

Telegram support bot

Prequesites:

• docker, docker-compose

Install:

1. Create new bot token using Telegram Bot Father, insert into bot.yaml

2. Create support chat only for agents, insert invite link to bot.yaml

3. Fill your support Agents data in bot.yaml

    bot_token: ""
    bot_company_name: CORP
    delivery_ratelimit: 30 // telegram limits up to 30 messages/sec from one bot
    support_chatid: -297716605
    support_link: https://t.me/joinchat/... // Options->Manage group->Create invite link
    default_sla: 12 # default sla in hours
    agents:
        - name: "agent1"
          chatid: *your_userid_here*
        - name: "agent2"
          chatid: *your_userid_here*
    db:
        # postgres connection settings (container)
        host: "127.0.0.1"
        dbname: "tgsup"
        user: "tgsup"
        password: "tgsup"
        sslmode: "disable"
        migrate: true
   

4. Run (you need docker, docker-compose installed)

   docker-compose build && docker-compose up
   

Usage manual

User

On the first start conversation will be automatically created and you will be registered

Enter short description of your problem, then feel free describe it precisely and attach any files

When problem is solved, press /current to end conversation

After that you may open another conversation using /active or select any resolved/closed conversation using /history

Agent

On the first start use invite link to join support chat

From agent point of view you can participate in any conversation, one at a time, using /active, or review/reopen conversation from /history

Use /search any text or caption of msg when talking to bot to reveal helpful solutions from your history

All conversation data stored in database before forwarding to Telegram, and will be restored when you join conversation, so in case of history glitch you will not lose any of your data

Support chat

In support chat one can see notifications of users and agents actions

Use /list in support chat to see relevant conversations, ordered by time spent

E2E Testing

    go test ./...

TODO

Software is only alpha for now, don't use it for production yet

• Refactor transport, make it testable
• Create rate system, agent can sent messages with inline rate keyboard, conversation will end with this keyboard too
• Improve search command with filters
• Create support statistics (support chat)