One of my favorite things about working at HubSpot is everyone's great sense of humor. The developer chatroom chronicles most of our zany antics, and it's facilitated by our very own robotic butler, Wooster Bot (named after Bertie Wooster, and based off of GitHub's Hubot).
#These bros want a chat bot...
Wooster is pretty smart; we've taught him how to evaulate mathematical expressions, translate text between languages, kick off new builds, generate a whole slew of different meme photos, and even tell us what today's specials are at the deli across the street. Any employee can add more functionality to Wooster by writing a node.js script. What else can we do with Wooster?
##JFDI
David Cancel, our fearless leader Chief Product Officer, is known for his poignant, humorous, and often quotable hashtags:
.@antrod lots of books and events cause a new sucker is born every day. #FAKERZ
David Cancel (@dcancel) April 9, 2012
Boston Startupers beware. #GeekswithGuns is back. #actlikeyouknow
David Cancel (@dcancel) April 5, 2012
The @HubSpot sales team is making it rain like a Lil Wayne video today! #majorleagues
David Cancel (@dcancel) March 30, 2012
So we figured, let's harness this corpus for some laughs.
#A man's gotta have a code.
We can implement this easily in CoffeeScript:
module.exports = (robot) -> # twitter API url url = 'https://api.twitter.com/1/statuses/' + 'user_timeline.json?include_entities=true' + '&screen_name=dcancel&count=199' # respond when someone says "@wooster dc" robot.response /dc$/i, (msg) -> # hit twitter API msg.http(url).get() (err, res, body) -> # catch HTTP errors if err return # parse response JSON data = JSON.parse body # catch API errors if data.error msg.send "Error: " + data.error return # get list of hashtags hashtags = tweet.entities.hashtags for tweet in data # flatten hashtag list hashtags = [].concat.apply [], hashtags # respond with a random hashtag if hashtags.length > 0 index = Math.floor(Math.random() * hashtags.length) msg.send '#' + hashtags[index].textOur final implementation makes use of Redis as a backing store to minimize how often we hit the Twitter API, but this was left out to improve clarity.
#Hello World
After deploying to our new Wooster script, we're in serendipitous hashtag heaven!
