[{"id":"cG9zdDoxNjc=","slug":"telegram-bots-are-easy","title":"Telegram bots are easy","content":"\n

Recently I’ve been frustrated with mobile stock apps so I started thinking about solutions. a whole app would have been an unnecessary overkill, after all I just want to get market data about my stocks with some indicators to go along. A perfect use case for a telegram bot!

\n\n\n\n

Today we will have a look at how telegram bots are written and show some base concepts around the platform. We are going to leverage python-telegram-bot (latest docs).

\n\n\n\n

What we will build today:

\n\n\n\n

A very simple bot that just replies a picture of a doggo whenever we send /woof.
A bot that has persistence and will save a list of items we can add.
(bonus round) Dockerize the bot for running it everywhere.

\n\n\n\n

\"\" — Photo by Arseny Togulev on Unsplash

\n\n\n\n

Here we go 🤖

\n\n\n\n

The first thing we need is a bot. To register one you need to talk to the @BotFather (with your normal telegram account). Then simply enter the /newbot command. You will be asked the name and the @handle for your bot. The handle need to end in bot.
After that he will answer with your access token.

\n\n\n\n

\"\" — Response after successful bot registration

\n\n\n\n

How do bots work?

\n\n\n\n

\"\" — very badly drawn representation of the bot

\n\n\n\n

In its most simple form bots are simple programs that listen to commands send by a user. The cannot initiate a conversation, a user must always message the bot first. After initial contact, the bot can send as many messages as he wants.

\n\n\n\n

Your code connects to the telegram servers and gets notified if new messages/commands have been sent. Then based on the user input your bot can reply with a message, audio, photo or use external services like sending a slack notification, whatever, there are no real limits.

\n\n\n\n

When the user firsts contacts a bot telegram automatically sends a /start command which you can use to initialise the bot. When a bot is deleted by a user the corresponding /stop command is issued.

\n\n\n\n

Doggo Bot

\n\n\n\n

We are going to create a virtual environment for the project and install the dependencies. Our starting point will be the basic bot skeleton from which we will fill out the functionality.

\n\n\n\n

Note: we need version 12 of python-telegram-bot, as the API changed quite a bit in the previous releases.

\n\n\n\n

python3 -m venv env\nsource env/bin/activate\npip install python-telegram-bot==12.0.0b1

\n\n\n\n

from telegram.ext import Updater, CommandHandler\n\nTOKEN = 'myTelegramBotToken'\n\ndef main():\n    updater = Updater(TOKEN, use_context=True)\n    dp = updater.dispatcher\n\n    updater.start_polling()\n    updater.idle()\n\n\nif __name__ == '__main__':\n    main()

\n\n\n\n

Great, now lets write a function to respond to a simple /ping message.

\n\n\n\n

// ...\n\ndef pong(update, context):\n    update.message.reply_text('Pong')\n\n\ndef main():\n    updater = Updater(TOKEN, use_context=True)\n    dp = updater.dispatcher\n\n    dp.add_handler(CommandHandler('ping', pong))\n\n    updater.start_polling()\n    updater.idle()\n\n// ...

\n\n\n\n

When we now send /ping to the bot we get a Pong back. Neat!

\n\n\n\n

\n\n\n\n

But we want doggos! 🐶

\n\n\n\n

def get_image_url():\n    allowed_extension = ['jpg', 'jpeg', 'png']\n    while True:\n        url = requests.get('https://random.dog/woof.json').json()['url']\n        file_extension = re.search(\"([^.]*)$\", url).group(1).lower()\n        if file_extension in allowed_extension:\n            break\n    return url\n\n\ndef woof(update, context):\n    update.message.reply_photo(photo=get_image_url())\n\n\ndef main():\n    //...\n\n    dp.add_handler(CommandHandler('ping', pong))\n    dp.add_handler(CommandHandler('woof', woof))\n\n    //...

\n\n\n\n

We define a second command that listens to the woof word. Then we have a little helper function that retrieves a url. That url is then passed to the reply_photo method and we are done!

\n\n\n\n

We could also pass a bite-like object to the photo parameter instead of a string containing the url if we would have an actual photo.

\n\n\n\n

\n\n\n\n

Persistence Bot

\n\n\n\n

The doggo bot is a good hello world example, but without persistence a lot of bots would not work. If you want to save something you will need to have some kind of user base storage.

\n\n\n\n

For this we will implement a little bot that saved a list of items that a user can add/delete. A bit like a shopping list if you want.

\n\n\n\n

Commands

\n\n\n\n

Add /add avocados
Delete /delete avocados
List all /all
Delete all /clear

\n\n\n\n

First the basic setup again. We will import PicklePersistence which uses the python pickle object to save the data on your drive. We need to initialise it and pass it to the Updater class.

\n\n\n\n

from telegram.ext import Updater, CommandHandler, CallbackContext, PicklePersistence\nfrom telegram import Update\n\nTOKEN = 'myTelegramBotToken'\n\npersistence = PicklePersistence('./db')\n\n\ndef main():\n    print('Started 🚀')\n    updater = Updater(TOKEN, use_context=True, persistence=persistence)\n    dp = updater.dispatcher\n\n    updater.start_polling()\n    updater.idle()\n\n\nif __name__ == '__main__':\n    main()

\n\n\n\n

Before we beginn lets define 2 helper functions:

\n\n\n\n

def get_list(context: CallbackContext) -> dict:\n    return context.user_data.setdefault('list', {})

\n\n\n\n

get_list will return the object that is unique to every user. The python library generously offers a user_data object inside of the context parameter that gets passed to each handler. We use setdefault in order to get the already present dict of the user or set it to an empty one if undefined.

\n\n\n\n

def parse_command(update: Update) -> (str, str):\n    key, value = update.message.text.split(' ', 1)\n    return key, value

\n\n\n\n

parse_command simply helps us interpret the message of the user. If a user sends us /add avocado we only want the avocado part. We simply split at the first space.

\n\n\n\n

Lets finally add all the methods.

\n\n\n\n

//...\n\ndef get_list(context: CallbackContext) -> list:\n    return context.user_data.setdefault('list', [])\n\n\ndef parse_command(update: Update) -> (str, str):\n    key, value = update.message.text.split(' ', 1)\n    return key, value\n\n\ndef list_add(update: Update, context: CallbackContext):\n    key, value = parse_command(update)\n    get_list(context).append(value)\n    update.message.reply_text('Saved 💾')\n\n\ndef list_delete(update: Update, context: CallbackContext):\n    key, value = parse_command(update)\n    get_list(context).remove(value)\n    update.message.reply_text('Deleted 🗑')\n\n\ndef list_all(update: Update, context: CallbackContext):\n    items = get_list(context)\n    update.message.reply_text('\\n'.join(items) if len(items) > 0 else 'List empty 😢')\n\n\ndef list_clear(update: Update, context: CallbackContext):\n    get_list(context).clear()\n    update.message.reply_text('Cleared 🧼')\n\n\ndef main():\n    //...\n    \n    dp.add_handler(CommandHandler('add', list_add))\n    dp.add_handler(CommandHandler('delete', list_delete))\n    dp.add_handler(CommandHandler('all', list_all))\n    dp.add_handler(CommandHandler('clear', list_clear))\n    \n    //...

\n\n\n\n

\n\n\n\n

I believe telegram bots can be a great alternative if a full fledged app is an overkill. There are a lot of featured that where not mentioned here, especially custom keyboards.

\n\n\n\n

Dockerize it

\n\n\n\n

# requirements.txt\n\npython-telegram-bot==12.0.0b1

\n\n\n\n

# Dockerfile\n\nFROM python:3.7-slim\nWORKDIR /app\n\nCOPY requirements.txt .\nRUN pip install -r requirements.txt && rm -rf /root/.cache\n\nCOPY src .\nCMD [\"python\", \"-u\", \"/app/Mercatus.py\"]

\n\n\n\n

# docker-compose.yml\n\nversion: '3.6'\n\nservices:\n  bot:\n    build: .\n    restart: unless-stopped\n    command: [\"python\", \"-u\", \"/app/Mercatus.py\"]\n    volumes:\n      - ./data.db:/app/data.db

\n\n\n\n

Thats it! I hope you enjoyed the guide. 👋

\n","status":"publish","date":"2019-06-24T17:31:52","modified":"2020-09-22T20:47:13","tags":{"nodes":[{"id":"dGVybTo1","slug":"automatization","name":"automatization","count":1},{"id":"dGVybTo3","slug":"bot","name":"bot","count":1},{"id":"dGVybTo0Ng==","slug":"python","name":"python","count":3},{"id":"dGVybTo1Mw==","slug":"telegram","name":"telegram","count":1}]},"post":{"featured":{"srcSet":"https://api.nicco.io/wp-content/uploads/2019/06/arseny-togulev-1513013-unsplash-300x169.jpg 300w, https://api.nicco.io/wp-content/uploads/2019/06/arseny-togulev-1513013-unsplash-1024x576.jpg 1024w, https://api.nicco.io/wp-content/uploads/2019/06/arseny-togulev-1513013-unsplash-150x84.jpg 150w, https://api.nicco.io/wp-content/uploads/2019/06/arseny-togulev-1513013-unsplash-768x432.jpg 768w, https://api.nicco.io/wp-content/uploads/2019/06/arseny-togulev-1513013-unsplash-1536x864.jpg 1536w, https://api.nicco.io/wp-content/uploads/2019/06/arseny-togulev-1513013-unsplash-2048x1152.jpg 2048w, https://api.nicco.io/wp-content/uploads/2019/06/arseny-togulev-1513013-unsplash-500x281.jpg 500w, https://api.nicco.io/wp-content/uploads/2019/06/arseny-togulev-1513013-unsplash-800x450.jpg 800w, https://api.nicco.io/wp-content/uploads/2019/06/arseny-togulev-1513013-unsplash-1200x675.jpg 1200w, https://api.nicco.io/wp-content/uploads/2019/06/arseny-togulev-1513013-unsplash-1500x844.jpg 1500w, https://api.nicco.io/wp-content/uploads/2019/06/arseny-togulev-1513013-unsplash-2000x1125.jpg 2000w","altText":"","sourceUrl":"https://api.nicco.io/wp-content/uploads/2019/06/arseny-togulev-1513013-unsplash-scaled.jpg"}}}]