2Ze.us Updates

There has been quite a bit of recent activity in my 2ze.us project since I first released it nearly a year ago. My intent was not to become a competitor with bit.ly, is.gd, or anyone else in the URL-shortening arena. I created the site as a way for me to learn more about Google's AppEngine. It didn't take very long to get it up and running, and it seemed to work fairly well.

AppEngine and Extensions

I was able to basically leave the site alone on AppEngine for several months--through about September 2009. In that time, I came up with a Firefox extension to make its use more convenient.

The extension allows you to quickly get a shortened URL for the page you're currently looking at, and a couple of context menu items let you get a short URL for things like specific images on a page. Also included in the extension is a preview for 2ze.us links. The preview can tell you the title and domain of the link's target. It can tell you how much smaller the 2ze.us URL is compared to the full URL. Finally, it displays how many times that particular 2ze.us link has been clicked.

That as all fine and dandy. It was the second Firefox extension I had ever written, and it's still running strong. In June or July of 2009, I started working on a little program to make it easier for me to interact with Twitter the way I wanted to. This was a great opportunity for me to incorporate 2ze.us into the application so any URL I wanted to post to Twitter would automatically be shortened for me, using my own shortener.

Porting to WebFaction And PHP

Anyway, around the end of September 2009, I noticed that there were a lot of problems with 2ze.us. It was slow and sometimes completely unresponsive. Certain URLs would redirect to their full URLs, while others wouldn't. The Firefox extension stopped working nicely. Oh yeah, and AppEngine rolled back to a previous revision of the code without me telling it to. That's when everything just died. It didn't take long for me to decide to migrate my project from AppEngine onto my awesome WebFaction hosting.

At this point, I was faced with a small dilemma: keep the code in Python, or port it to PHP. I opted to port it over to PHP, because I didn't want all of the overhead of a full Django instance for a site that needed to be very zippy. And I was unacquainted with other Python options.

By early October 2009, I had managed to turn the project into a PHP beast, running on Apache. It was a lot more responsive than AppEngine ever let 2ze.us be. There were a few bumps along the road, what with the extension and Twitter client relying on various parts of the site. Eventually it got to a point where I could just let it sit and work.

Chromium Extension

Sometime around the end of December, I decided to write another extension for 2ze.us, only for Google Chrome and Chromium this time. This extension isn't quite as feature-packed as its Firefox brother, but it gets the job done.


Shortly after "completing" the Chromium extension, I had what seemed like a pretty original idea. Who knows if it really is, but I still haven't seen another tool quite like the one that I made as a result of this idea. I thought, "Now, why should I need to install an extension in each Web browser I use on each computer I use? Is there a better way?"

The answer came quickly: a standalone, desktop application. Write one program that handles shortening URLs for you. My laziness told me to make a program that monitors your system clipboard for URLs. If a URL is detected, try to shorten it, and update the clipboard contents in place. Boom. Done. All extensions become useless beyond things like the URL preview (which is very useful, imo).

The next question I asked was, "Do I make it platform-dependent? Should I stick it to the majority of computer users and write my tool for Linux only? For OSX only? For, uh... Windows only?" Again, an easy question to answer. Support them all or don't even bother writing the application.

A week's worth of midnight hacking saw the birth of Clip2Zeus 1.0a. It's a cross-platform compatible desktop application that does exactly what I just mentioned. When it's running and detects a URL on your system clipboard, it will try to shorten it and update it in your clipboard. If you copy a block of text, the application will only modify the URLs in that block of text--meaning the block of text will still be in your clipboard, but it will have shorter URLs.

I use the program every day at work (on OSX). It's been very fun for me to see a short URL any time I copy a nasty URL to my clipboard. Imagine that; I'm a big fan of my own work...


Lately, I've noticed that the site was getting kind of slow again. Sometimes it would take several seconds for Clip2Zeus to shorten URLs in my clipboard, when it was normally instantaneous. Every once in a while, Clip2Zeus would completely fail to connect to the website.

One of my friends has asked me a lot of questions about the Tornado framework in the past months. I had read a few things about Tornado when it was open-sourced last year, but I didn't really feel the need to dabble with it. These questions prompted me to tinker a little.

Last night I re-ported 2ze.us to Python, using the Tornado framework this time. So far I'm very impressed with its responsiveness. The framework offers a lot of neat little utilities, and it is very fast (as reported by dozens of other reputable sources).

On top of the speed increase that came with the transition to Tornado, my RAM usage on WebFaction has come down by nearly 100MB. Just by turning off the one Apache-backed website. Now I'm nowhere near my RAM cap! Wahoo!!

Enough rambling. Like I said at the beginning of this article, a lot has been happening with this project in the past year. I didn't even think about all of the time I put into projects related to my simple little side project. Looking back, I'm quite satisfied with how things have unfolded.


Here are some simple statistics for 2ze.us. Since March 2009...

  • 5,252 URLs have been shortened using 2ze.us
  • 2ze.us links have been clicked 198,267 times
  • 315,951 URL characters have been turned into 11,532 characters

In April 2009...

  • 217 URLs were shortened
  • 2ze.us links were clicked 617 times

In February 2010...

  • 1,182 URLs were shortened
  • 2ze.us links were clicked 32,830 times

Not too shabby for a side project.

My Fedora 11 Adventures: Part VI

Folks, I cannot take this any longer. I've had Fedora 11 installed on my computer for 5 days now. That is close enough to a week for me. There simply is not enough about Fedora right now to keep me using it. Perhaps the next release will be better for me. I honestly hope so.

To be perfectly honest, I enjoyed most of the Fedora experience these past few days. I was thoroughly impressed with the speed and memory usage in Fedora compared to Jaunty. When I mentioned that on Twitter the other day, one fellow asked if the two systems were running the exact same software. His train of thought seemed to be that you can't really compare two different distros for speed or memory usage unless they run the exact same software at the time of the sample.

My response to that is that it doesn't matter to me in this particular case. I was comparing the general performance of both distros using their "stock" configuration. You can customize a distro however you'd like, and, in the end, that's where you'll probably find the most performance gains in any system.

But performance out of the box is important to me. I'll just leave it at that.

As I write this, I'm creating an ISO of slackware-current (as of midnight MST) so I can see what KDE 4 is like on a real distribution. Heh. This oughta be fun. Anyway, I truly hope that the next release of Fedora will hold my attention for a bit longer.

Checking In

I suppose I should update everyone out there about what I've been up to lately. It seems strange to me that I post article much less frequently now than I did when I was a full-time university student. You'd think I'd have a whole lot more time to blog about whatever I've been working on. I suppose I do indeed have that time, it's just that I usually like to wait until my projects are "ready" for the public before I write about them.

The biggest reason I haven't posted much of anything lately is a small Twitter client I've been working on. Its purpose is to be a simple, out-of-the-way Twitter client that works equally well on Windows, Linux, and OSX. The application is written in Python and wxPython, and it has been coming along quite well. It works great in Linux (in GNOME and KDE at least), but Windows and OSX have issues with windows stealing focus when I don't want them to. I'm still trying to figure it out--any advice would be greatly appreciated.

Chirpy currently does nothing more than check your Twitter accounts for updates periodically. It notifies you of new updates using blinking buttons (which can be configured to not blink). I think the interface is pretty nice and easy to use, but I am its developer so it's only proper that I think that way.

Anyway, that project has been sucking up a lot of my free time. It's been frustrating as I build it in Linux only to find that Windows and OSX both act stupidly when I go to test it. That frustration inspired me to tinker with a different approach to a Twitter client. I began fooling around with it last night, and I think the idea has turned out to be more useful than Chripy is after a month of development!

I'm calling this new project "Tim", which is short for "Twitter IM". This one also periodically checks your Twitter account(s) for updates (of course). However, Tim will send any Twitter updates to any Jabber-enabled instant messenger client that you are signed into. If you're like me, you have Google Talk open most of the day, so you can just have Twitter updates go straight there! You can also post updates to Twitter using your Jabber instant messenger when Tim is running by simply sending a message back!!

The really neat stuff comes in when you start to consider the commands that I've added to Tim tonight. I've made it possible for you to filter out certain hashtags, follow/unfollow users, and specify from which Twitter account to post updates (when you have multiple accounts enabled). I hate all of those #FollowFriday tweets... they drive me crazy. So all I have to do is type ./filter followfriday and no tweet that contains #FollowFriday will be sent to my Jabber client. I love it.

More commands are on the way. Also on the way is a friendly interface for configuring Tim. Getting it up and running the first time is... a little less than pleasant :) Once you have it configured it seems to work pretty well though.

If you're interested in trying it out, just head on over to the project's page (http://bitbucket.org/codekoala/twitter-im/). Windows users can download an installer from the Downloads tab. I plan on putting up a DMG a little later tonight for OSX users. Linux users can download the .tar.gz file and install the normal Python way :) Enjoy!

Update: The DMG for OSX is a little bigger than I thought it would be, so I won't be hosting it on bitbucket. Instead, you can download it from my server.

Don't forget to read the README !!!

AES Encryption in Python Using PyCrypto


Please do not mistake this article for anything more than what it is: my feeble attempt at learning how to use PyCrypto. If you need to use encryption in your project, do not rely on this code. It is bad. It will haunt you. And some cute creature somewhere will surely die a painful death. Don't let that happen.

If you want encryption in Python, you may be interested in these libraries:

I spent a little bit of time last night and this morning trying to find some examples for AES encryption using Python and PyCrypto. To my surprise, I had quite a difficult time finding an example of how to do it! I posted a message on Twitter asking for any solid examples, but people mostly just responded with things I had seen before--the libraries that do the encryption, not examples for how to use the libraries.

It wasn't long after that when I just decided to tackle the problem myself. My solution ended up being pretty simple (which is probably why there weren't any solid examples for me to find). However, out of respect for those out there who might still be looking for a solid example, here is my solution:

#!/usr/bin/env python

from Crypto.Cipher import AES
import base64
import os

# the block size for the cipher object; must be 16 per FIPS-197

# the character used for padding--with a block cipher such as AES, the value
# you encrypt must be a multiple of BLOCK_SIZE in length.  This character is
# used to ensure that your value is always a multiple of BLOCK_SIZE

# one-liner to sufficiently pad the text to be encrypted
pad = lambda s: s + (BLOCK_SIZE - len(s) % BLOCK_SIZE) * PADDING

# one-liners to encrypt/encode and decrypt/decode a string
# encrypt with AES, encode with base64
EncodeAES = lambda c, s: base64.b64encode(c.encrypt(pad(s)))
DecodeAES = lambda c, e: c.decrypt(base64.b64decode(e)).rstrip(PADDING)

# generate a random secret key
secret = os.urandom(BLOCK_SIZE)

# create a cipher object using the random secret
cipher = AES.new(secret)

# encode a string
encoded = EncodeAES(cipher, 'password')
print 'Encrypted string:', encoded

# decode the encoded string
decoded = DecodeAES(cipher, encoded)
print 'Decrypted string:', decoded

Edit: thanks to John and Kaso for their suggestions, though John's didn't seem to work for me (?)

Edit 2015.12.14: thanks to Stephen for pointing out that the block size for AES is always 16, and the key size can be 16, 24, or 32. See FIPS-197 for more details.

If you plan to use this script, you'll need to have PyCrypto installed on your computer. I have had a difficult time finding this for Windows in the past, so I will mirror the installer that I found over here: http://jintoreedwine.wordpress.com/2008/07/20/python-25-and-encryption-pycrypto-under-windows/. I haven't tried it on Mac OS X yet, but it should be fairly simple to install it. Same goes for Linux.

The output of the script should always change with each execution thanks to the random secret key. Here's some sample output:

$ python aes_encryption.py
Encrypted string: aPCQ8v9WzLM/JusrJPS19K8uUA/34Xiu/ZR+arzl1oM=
Decrypted string: password

$ python aes_encryption.py
Encrypted string: F0cp4hMk8RXjcww270leHnigH++yqysIyPy8Em/qEbI=
Decrypted string: password

$ python aes_encryption.py
Encrypted string: 7gH2QCIPOxXVBjTXrMmdgU2l7Iku5Lch5jpG9OScGZw=
Decrypted string: password

$ python aes_encryption.py
Encrypted string: oJUq0/XHdmYgC3ILgFgF6Tpuo8ZhoEHN9wmnuYvV58Y=
Decrypted string: password

If the comments in the script aren't explanatory enough, please comment and ask for clarification. I will offer any that I am capable of, and I invite others to do the same.

Send E-mails When You Get @replies On Twitter

I just had a buddy of mine ask me to write a script that would send an e-mail to you whenever you get an "@reply" on Twitter. I've recently been doing some work on a Twitter application, so I feel relatively comfortable with the python-twitter project to access Twitter. It didn't take very long to come up with this script, and it appears to work fine for us (using a cronjob to run the script periodically).

I thought others on the Internets might enjoy the script as well, so here it is!

#!/usr/bin/env python
# -*- coding: utf-8 -*-

A simple script to check your Twitter account for @replies and send you an email
if it finds any new ones since the last time it checked.  It was developed using
python-twitter 0.5 and Python 2.5.  It has been tested on Linux only, but it
should work fine on other platforms as well.  This script is intended to be
executed by a cron manager or scheduled task manager.

Copyright (c) 2009, Josh VanderLinden
All rights reserved.

Redistribution and use in source and binary forms, with or without modification,
are permitted provided that the following conditions are met:

- Redistributions of source code must retain the above copyright notice, this
list of conditions and the following disclaimer.
- Redistributions in binary form must reproduce the above copyright notice, this
list of conditions and the following disclaimer in the documentation and/or
other materials provided with the distribution.
- Neither the name of the organization nor the names of its contributors may
be used to endorse or promote products derived from this software without
specific prior written permission.


import twitter
import ConfigParser
import os
import sys
from datetime import datetime
import smtplib
from email.MIMEMultipart import MIMEMultipart
from email.MIMEText import MIMEText
from email.Utils import formatdate

# get the user's "home" directory
DIRNAME = os.path.expanduser('~')
CONFIG = os.path.join(DIRNAME, '.twitter_email_replies.conf')
FORMAT = '%a %b %d %H:%M:%S +0000 %Y'
REPLY_TEMPLATE = """%(author)s said: %(text)s
Posted on %(created_at)s
Go to http://twitter.com/home?status=@%(screen_name)s%%20&in_reply_to_status_id=%(id)s&in_reply_to=%(screen_name)s to post a reply

# sections
AUTH = 'credentials'
EXEC = 'exec_info'
EMAIL = 'email_info'

# make the code a bit "cleaner"
O = lambda s: sys.stdout.write(s + '\n')
E = lambda s: sys.stderr.write(s + '\n')
str2dt = lambda s: datetime.strptime(s, FORMAT)

def get_dict(status):
    my_dict = status.AsDict()
    my_dict['screen_name'] = my_dict['user']['screen_name']
    my_dict['author'] = my_dict['user']['name']
    return my_dict

def main():
    O('Reading configuration from %s' % CONFIG)
    parser = ConfigParser.SafeConfigParser()
    config = parser.read(CONFIG)

    # make sure we have the proper sections
    if not parser.has_section(AUTH): parser.add_section(AUTH)
    if not parser.has_section(EMAIL): parser.add_section(EMAIL)
    if not parser.has_section(EXEC): parser.add_section(EXEC)

        # get some useful settings from the configuration file
        username = parser.get(AUTH, 'username')
        password = parser.get(AUTH, 'password')

        to_address = parser.get(EMAIL, 'to_address')
        from_address = parser.get(EMAIL, 'from_address')
        smtp_server = parser.get(EMAIL, 'smtp_server')
        smtp_user = parser.get(EMAIL, 'smtp_user')
        smtp_pass = parser.get(EMAIL, 'smtp_pass')

        if '' in [username, password, to_address, from_address, smtp_server]:
            raise Exception('Not configured')
    except Exception:
        E('Please configure your credentials and e-mail information in %s!' % CONFIG)

        # create some placeholders in the configuration file to make it easier
        sections = {
            AUTH: ('username', 'password'),
            EMAIL: ('to_address', 'from_address', 'smtp_server', 'smtp_user', 'smtp_pass')

        for section in sections.keys():
            for opt in sections[section]:
                if not parser.has_option(section, opt):
                    parser.set(section, opt, '')
        # determine the last time we checked for replies
            last_check = str2dt(parser.get(EXEC, 'last_run'))
        except ConfigParser.NoOptionError:
            last_check = datetime.utcnow()
        last_check_str = last_check.strftime(FORMAT)

        info = 'Fetching updates for %s since %s' % (username,

        # attempt to connect to Twitter
        api = twitter.Api(username=username, password=password)

        # not using the `since` parameter for more backward-compatibility
        timeline = api.GetReplies()
        new_replies = []
        for reply in timeline:
            post_time = str2dt(reply.GetCreatedAt())
            if post_time > last_check:

        count = len(new_replies)
        if count:
            # send out an email for this user
            O('Found %i new replies... sending e-mail to %s' % (count, to_address))
            reply_list = '\n\n'.join([REPLY_TEMPLATE % get_dict(r) for r in new_replies])
            is_are = 'is'
            plural = 'y'
            if count != 1:
                is_are = 'are'
                plural = 'ies'

            params = {
                'is_are': is_are,
                'count': count,
                'replies': plural,
                'username': username,
                'reply_list': reply_list,
                'last_check': last_check_str

            text = """There %(is_are)s %(count)i new @repl%(replies)s for %(username)s on Twitter since %(last_check)s:

%(reply_list)s""" % params

            # compose the e-mail
            msg = MIMEMultipart()
            msg['From'] = from_address
            msg['To'] = to_address
            msg['Date'] = formatdate(localtime=True)
            msg['Subject'] = 'New @Replies for %s' % username

            # try to send the e-mail message out
            email = smtplib.SMTP(smtp_server)
            if smtp_user and smtp_pass:
                email.login(smtp_user, smtp_pass)

        # save the current time so we know where to pick up next time
        parser.set(EXEC, 'last_run', datetime.utcnow().strftime(FORMAT))

    # write the config
    O('Saving settings...')
    out = open(CONFIG, 'wb')

if __name__ == '__main__':

Feel free to copy this script and modify it to your desires. Also, please comment if you have issues using it.