Easy! Implement a Twitter bot that runs on Heroku in Python

I will show you how to implement a Twitter bot that runs on Heroku in Python. I think it's exhausted, but it's the 2016 version.

The solution is to deploy a Python script on Heroku and use Heroku Scheduler to run the script on a regular basis.

The source code of the completed version is below. https://github.com/k-enomoto/minimum_twitter_bot

What to explain in this article

--Minimum source code for Twitter Bot --Deploy to Heroku --Heroku Scheduler settings

Not explained in this article

--How to use Heroku --How to use pip --How to create a Twitter API account

There are other detailed articles for each, so please refer to them.

Library preparation

Install the library you want to use.

$ pip install python-twitter
$ pip install bottle

We need to tell Heroku which library to use, so create requirements.txt.

$ pip freeze -l > requirements.txt

Maybe it will be like this.

bottle==0.12.9
future==0.15.2
oauthlib==1.1.2
python-twitter==3.1
requests==2.10.0
requests-oauthlib==0.6.2

Code to access the Twitter API

Get the Twitter API consumer_key, consumer_secret, ʻaccess_token_key, ʻaccess_token_secret in advance. Please refer to the following article for the acquisition method. Tweet using Twitter API in Ruby Also, each key is set in an environment variable from the viewpoint of security. How to set environment variables on Heroku will be described later.

Please feel free to arrange the content of the tweet as soon as you take it from the Web API.

tweet.py

# -*- coding: utf-8 -*-

import os
from datetime import datetime

import twitter


api = twitter.Api(consumer_key=os.environ["CONSUMER_KEY"],
                  consumer_secret=os.environ["CONSUMER_SECRET"],
                  access_token_key=os.environ["ACCESS_TOKEN_KEY"],
                  access_token_secret=os.environ["ACCESS_TOKEN_SECRET"]
                  )
api.PostUpdate("system time is %s" % datetime.now())

Web application code (dummy)

This is the point of this time. Prepare a dummy web app to keep the process running on Heroku. In this article, we used the lightweight web framework bottle. The sample code in bottle documentation is used as is.

index.py

# -*- coding: utf-8 -*-

import os
from bottle import route, run


@route("/")
def hello_world():
    return "" #Whatever you return here

run(host="0.0.0.0", port=int(os.environ.get("PORT", 5000)))

Procfile The Procfile contains the commands to be executed in the Heroku container. The command to start the above dummy web application is described.

Procfile

web: python index.py

runtime.txt This is point 2. Create runtime.txt and describe the runtime to use.

runtime.txt

python-3.5.2

Deploy

Now you have all the files you need.

index.py
tweet.py
requirements.txt
runtime.txt
Procfile

Files must be managed in Git in order to deploy to Heroku. Commit to Git.

git add .
git commit -m "Initial commit"

Deploy to Heroku. Also, the environment variables mentioned above are set here. Replace your_cosumer_key, your_consumer_secret, your_access_token_key, and your_access_token_secret with your own values.

$ heroku create --stack cedar
$ git push heroku master
$ heroku config:set CONSUMER_KEY=your_cosumer_key CONSUMER_SECRET=your_consumer_secret ACCESS_TOKEN_KEY=your_access_token_key ACCESS_TOKEN_SECRET=your_access_token_secret

Make sure that you can deploy properly.

$　heroku logs

Registering a scheduler task

Finally, register the task with Heroku Scheduler. Enter the following command to bring up the scheduler registration page.

$ heroku addons:create scheduler:standard
$ heroku addons:open scheduler

Set the command to be executed by the scheduler and the frequency. Let's set up a Python script to tweet.

$ python tweet.py

This completes the Twitter bot.