Steps to create a Twitter bot with python

Date of writing: 2015.08.09

Thank you for always referring to your blog, Qita, Github, etc.

1. Preparation

1. Access Twitter App
2. Create a new app with [Create New App]
3. Get various keys

• Consumer Key(API Key)
• Consumer Secret(API Secret)
• Access Token
• Access Token Secret

• Access Level should be changed according to the application you want to create.

2. Handle Twitter with Python

The environment is Python 3.5.1. There are many libraries that work with Twitter on Python. I've used the following three.

• python-twitter
• Tweepy
• TwitterAPI

Mostly none

$ pip install twitter

I think you can usually install it like this.

Twitter bot

1. Regular tweet type
2. Real-time response type

I think there are two types of (though there may be others)

This time I would like to create these two types of bots.

1. Regular tweet type

import twitter

auth = twitter.OAuth(consumer_key="",
consumer_secret="",
token="",
token_secret="")

t = twitter.Twitter(auth=auth)

#Tweet only
status="Hello,World" #Tweet to post
t.statuses.update(status=status) #Post to Twitter

#Tweet with image
pic=""　#If you post an image, the image path
with open(pic,"rb") as image_file:
 image_data=image_file.read()
pic_upload = twitter.Twitter(domain='upload.twitter.com',auth=auth)
id_img1 = pic_upload.media.upload(media=image_data)["media_id_string"]
t.statuses.update(status=status,media_ids=",".join([id=img1]))

If it is a bot that regularly tweets images and words, it is ok if you modify this a little!

There are several ways to run a script regularly, but using crontab is the best (I personally hate the notation).

Crontab is a tool that executes commands regularly when the set time is reached. -How to write crontab

2. Real-time response type

Use Streaming API.

import twitter

auth = twitter.OAuth(consumer_key="",
consumer_secret="",
token="",
token=secret="")

t = twitter.Twitter(auth=auth)
#Use Userstream
t_userstream = twitter.TwitterStream(auth=auth,domain='userstream.twitter.com')

#Your timeline tweets and user information flow
for msg in twitter.userstream.user():
 print(msg)

In the program, in the contents of msg Various information such as user information and tweet information will flow one by one, so please process it so that you can get the information you want.

I will post only the ones I use often.

I wonder if I haven't used the others so far ..

For example, if you want to create a bot that will reply when a reply comes

.
.
for msg in twitter.userstream.user():
 #Reply to TL TWITTER_Branch if ID is included
 if msg['in_reply_to_screen_name']=="TWITTER_ID":
  tweet = "@"+msg['user']['screen_name']+" "+"Reply Thanks you"
  t.statuses.update(status=tweet)
  

It can be created like this (example)

For those who want to know in detail what kind of information can be obtained from the flowing information (msg) [Data structure when acquired by Twitter API](http://biokids.org/?%A4%C9%A4%D6%A4%AA%2FTwitter%A4%C7%CD%B7%A4%DC%A4% Please see A6% A1% AA% 2F% A5% C7% A1% BC% A5% BF% B9% BD% C2% A4).

I will write about tweepy at a later date.

Actually, there may be a more efficient operation method, but for now it looks like this.

update_name Feeling like playing with twitter profile settings on CUI

import tweepy

auth = OAuthHandler(consumer_key="",consumer_secret="")
auth.set_access_token(token="",token=secret="")
api = API(auth)
api.update_profile(name='New name')

For other update_profiles, see POST account / update_profile | Twitter Developers.

For me personally, tweepy is easier to see and use than python-twitter. API Reference — tweepy 3.5.0 documentation So if you think you can do this, go to the twitter official and the reference of each library !!

Site that I was allowed to refer to

-Kivantium activity diary Using Twitter with Python ~ Introduction of Tweepy ~