This example uses the tweepy library for Python, which you can install by running pip install -r requirements.txt. You can find the full documentation here.
This script assumes your API keys are stored in the following environment variables:
TWITTERBOT_CONSUMER_KEYTWITTERBOT_CONSUMER_SECRETTWITTERBOT_ACCESS_TOKENTWITTERBOT_ACCESS_SECRET
You can also modify this section to read your keys from a file, database, or some other source - just don't push them to your public repo!
After logging in, the script creates an array:
trigger_words = [
'@' + bot.screen_name
]Any strings you put into this array will be tracked by tweepy, which will trigger a function we're going to implement whenever a Tweet is posted containing any of those words. The bot's handle is included by default to respond to mentions.
The function that gets fired is Stream.on_status(), which by default just prints the Tweet to the console. Here we create a new class MyStream which inherits from Stream where we'll override that function.
It takes a single argument, status, which contains info about the Tweet. Important to us is the following:
status.text- the message textstatus.author.screen_name- the person's @username
The new function analyzes the Tweet text (in this case, checking if it contains "hi" or "hello") and creates a variable reply containing the reply, then sends the Tweet with update_status().