Agrothon

Introduction

• This project has three parts

  • The Agrothon-Client Module which will be running in Raspberry Pi
  • API Server
  • Telegram Bot

• API Server handles Everything, All the routes are shown below

  

• Telegram bot is just a frontend for the whole Project

• Agrothon-Client Sends Sensor data, Intruder images to the API Server which will be analysed there and Stored in the Database

Installation

• Via pip

  • Install Dependencies

    mkdir agrothon && cd agrothon
    sudo add-apt-repository ppa:deadsnakes/ppa
    sudo apt-get install python3.8 wget
    wget -q https://github.com/caddyserver/caddy/releases/download/v2.4.1/caddy_2.4.1_linux_amd64.tar.gz
    tar xzf caddy_2.4.1_linux_amd64.tar.gz
    rm -rf caddy_2.4.1_linux_amd64.tar.gz
    chmod a+x caddy
    mv caddy /usr/local/bin/caddy
    wget -q https://github.com/viswanathbalusu/Agrothon/raw/main/Caddyfile -O Caddyfile
    

  • First fill the variables in agrothon-sample.env and rename it to agrothon.env

    wget -q https://github.com/viswanathbalusu/Agrothon/raw/main/agrothon-sample.env -O agrothon.env
    

  • you can extend all tha variables from Base Config

  • Get the latest Release of data directory and untar in it the same directory where you placed agrothon.env

    wget -q https://github.com/viswanathbalusu/Agrothon/releases/latest/download/Agrothon-Data.tar.gz
    tar -xzf Agrothon-Data.tar.gz && rm -rf Agrothon-Data.tar.gz
    

  • Edit the $API_PORT in Caddyfile or Just Set $API_PORT in your Shell Environment.

  • Then Create a Virtual Environment (Optional but Recommended) and then install Agrothon with

    python3.8 -m virtualenv venv
    source venv/bin/activate
    pip install Agrothon
    

  • There are two commands in Agrothon

    • agroserver - Which actually starts the Uvicorn Server on a Unix Domain Socket at /usr/agrothon.sock, So you should use a Reverse proxy (Preferably Caddy)
    • agrobot - Which starts the telegram bot
    • agrothon - Starts Both agroserver and agrobot with caddy reverse proxy

• Via Docker

  • Download Docker compose and Map the ports according to your use
    • wget -q https://viswanathbalusu.github.io/Agrothon/docker-compose.yml
  • Download agrothon.env and Fill the Variables (can be extended from Base Config)
    • wget -q https://viswanathbalusu.github.io/Agrothon/agrothon-sample.env -O agrothon.env
  • Finally do docker-compose up it will pull the image from container registry and run the services

Variables in agrothon.env

* if there are Other ports too the Bot will Work but it may not be able to post images

Note

• We are using heavy modules like tensorflow and Yolov3, so make sure you have atleast 2.5GB RAM free to get this running
• Only Works in UNIX environments
• caddy,xz-utils, all OpenCV Dependencies must be installed if you are choosing non Docker method.
• Python3.8 is compulsory

• Telegram Bot Start

  

• Bot Commands

    start - start command
    field - get field status
    weather - get weather status
    rainfall - get predictions of rainfall for your region
    settings - change settings of the bot
    restart - restart the bot
    log - get the log files
    stats - get the server stats
    ping - check ping of server
    help - to get help message
  

  

• Object Detection

  

  

  

  

  

  

• Bot Commands Usage

  • Pump Prediction

    

  • Language Change

    

  • Complete info

    

  • Field Command

    

  • Help Command

    

  • Object Detection by sending an Image to Bot

    

  • Language Menu

    

  • Pump Keyboard

    

  • Pump on/off

    

  • Rainfall Prediction

    

  • Settings Keyboard

    

  • Server Stats

    

  • Weather Data