garyalan.net

I started off with an Ubuntu 10.04 box trying to install Holla. That didn’t work too well, I ran into troubles getting the right version of everything installed. Despite my view that arch linux shouldn’t be used in a production environment, I was pretty sure it would have up-to-date packages, so I spun up a vm of arch on my HP MicroServer running ESXi and went to town.

The instructions on holla’s site to install holla are as follows:

• bundle install
• rake db:migrate
• rails server thin
• start the Juggernaut server – http://github.com/maccman/juggernaut
• Now go to http://localhost:3000

Simple, right? Well it looks like we need juggernaut. Instructions to install juggernaut on juggernaut’s site:

$ rvm alias create default ruby-1.9.2-p290

Now, to install rubygems. I installed from source:

Now we’re starting to get somewhere! We have ruby installed! Now we need to install juggernaut, so onto install npm to install juggernaut with. I had to install this as root, even with sudo it wouldn’t install for some reason.

These guys have some tricky installers. That’s the second time we’ve ran a script we got through curl in a one-liner.

So we have npm. Time to install juggernaut.

Now, to install its gem:

Great! Now we can start installing holla (finally!). First we need to clone the repository, install Gem bundler then run ‘bundle install’. Note that you have to re-source your .bash_profile to be able to use bundler.

Well, that stinks. Fortunately, I found a fix.

Great. That’s done. Back to bundle install.

Whoo! It worked. Onto step 2 (ha ha ha…we’re just on step 2???). Using bundle exec didn’t fix rake for me, but the following did. I found the fix here.

Next, step 3:

Finally, starting juggernaut

Head out to your server’s hostname or IP on port 3000 on a browser (make sure you use Chrome or Safari and you’ll have to log in with twitter)! Then, hopefully, it works like mine! Add a channel to get started, then test it!

And there you go!That’s it!  I was rather surprised it was working. This was the fourth try of the weekend and I was about ready to throw in the towel. Good luck to you if you try it! It was a super fun project and hopefully it will work for you!

[this was originally published 18 January 2012]

Walking home from work one day, I thought it would be fun to write my very own twitter bot. For some reason, python seemed like an appropriate language to do this in, so I did a little research. My two main sources ended up being @the_shrinkbot and @grammer_man (warning, not safe for work language on grammer_man’s site).

I needed an idea. A twitter bot is great, but it has to do something. I decided that there must be people who need an amen, so, I made a new account: @getanamen, mostly to amuse myself and my roommate.

#!/usr/bin/env python

from twitter.api import Twitter, TwitterError
from twitter.oauth import OAuth, write_token_file, read_token_file
from twitter.oauth_dance import oauth_dance
from time import gmtime, strftime
from sys import stdout

import os
import time
import sys
import pickle
import random

# I honestly don't know whose keys these are
# I think they might be twitter for command line's?
CONSUMER_KEY='uS6hO2sV6tDKIOeVjhnFnQ'
CONSUMER_SECRET='MEYTOS97VvlHX7K1rwHPEqVpTSqZ71HtvoK4sVuYk'

# thanks to ventolin and his bot, grammer_man
# source: http://ventolin.org/2012/01/grammer_man-who-the-fuck-is-this-nigga-and-why-u-comin-at-me-like-that-hoeassnigga/
def check_if_done(id):
    '''
    Checks if a tweet has already been responded to
    '''
    if os.path.exists('done.pkl'):
        f = open('done.pkl', 'r')
        done = pickle.load(f)
        f.close()
        if id in done:
            return True
    return False

# also thanks to ventolin and grammer_man
def update_done(id):
    '''
    Updates a list of tweets that've been replied to
    '''
    if os.path.exists('done.pkl'):
        f = open('done.pkl', 'r')
        done = pickle.load(f)
        f.close()
    else:
        done = []

    done.append(id)

    f = open('done.pkl', 'w')
    pickle.dump(done, f)
    f.close()

# a lot of this code is thanks to Vivek Haldar and his bot
# doctor_eliza
# http://blog.vivekhaldar.com/post/2830035130/how-to-write-a-twitter-bot-in-python
if __name__ == '__main__':
    oauth_filename = os.environ.get('HOME', '') + os.sep + '.twitter_oauth'
    oauth_token, oauth_token_secret = read_token_file(oauth_filename)

    # vivek: We use two twitter clients, one to search, another to update. Just
    # easier that way...
    twitter = Twitter(domain='search.twitter.com')
    twitter.uriparts=()
    print '##### args = ', sys.argv
    # some setup....
    last_id_replied = ''
    counter  = 0
    if len(sys.argv) > 1:
        counter = int(sys.argv[1])

    # the so-called oauth dance
    poster = Twitter(
        auth=OAuth(oauth_token, oauth_token_secret, CONSUMER_KEY, CONSUMER_SECRET),
        secure=True,
        api_version='1',
        domain='api.twitter.com')

    # this is main loop of the program
    while True:
	# search for #amen
	print 'Searching...'
        results = twitter.search(q="#amen", since_id=last_id_replied)['results']

        if not results:
            print 'No results this time...'

        for result in results:
            # Remove my name from the question.
	    # this made more sense for doctor_eliza,
	    # but I'm keeping it in for the terminal
	    # output.
            request = result['text'].replace('@GetAnAmen', '')
            asker = result['from_user']
            id = str(result['id'])

            # this is my stuff -- formatting the termianl output
            # because I enjoy monitoring it...
	    myTime = strftime("%a %b %H:%M:%S", gmtime())
	    print myTime
            print "\t" + asker + ":"
	    print "\t  " + request

	    # my response
            response = "Amen!"

            # creating the reply message
            message = '@%s %s' % (asker, response)
            print '\t@GetAnAmen responds:'
	    print '\t  ' + message
            last_id_replied = id

	    # check if I've already replied to this message
            if not check_if_done(id):
                # post!
                poster.statuses.update(status=message)
                counter += 1
                # add who I tweeted at as a friend
		# poster.friendships.create(id=asker)
		# add to list of who I've replied to (pickle is pretty cool)
                update_done(last_id_replied)
                print '\tLast id replied = ', last_id_replied
                print '\t%s Done\n' % counter
            else:
                # output to the console
                print '                  !!!!! already done !!!!! '
                print '                  %s Done\n' % counter
            # don't want to flood the twitters too badly
            time.sleep(random.randint(1,2))

	    # Joel and I did the math, and we think tweeting 40 times then
            # sleeping for an hour is good. Maybe change it to 20 / 30 mins
            # or 10 / 15 mins in the future? vamos a ver...
            if counter > 39:
                print 'Done %s tweets, sleeping for 60 minutes'
                for i in range(0,60):
                    stdout.write("%d " % i)
                    stdout.flush()
                    time.sleep(60)
                counter = 0
                # break the for loop so I search for more results
                break

    print "This should never be outputted"

After many iterations and being banned from twitter a few times for posting too much (#oops), this is more or less “who” @getanamen is:

It’s fun to watch who retweets, mentions, and follows him. You can see what he’s saying here: @getanamen. I hope you enjoy @getanament.

He’s built using Python Twitter Tools (ptt) and right now, running on a virtual machine on my desktop.