Vuong Nguyen

Migrating to Django 1.5 and Custom User Model

Vuong Nguyen — Thu, 28 Mar 2013 00:00:00 -0400

With the recent Django 1.5 release, we now have the choice to create custom user model instead of the default user model that came with django.contrib.auth.

Two important use cases this release resolved were the ability to easily extend user profile and fix User.email field from having a hard coded value of max_length=75. This isn’t in compliant with RFC 5321, which establishes that an email can have max_length=254.

Here are a few quick steps to migrate your current users and profiles to take advantage of the new custom user model feature in Django 1.5 release.

Create a new Django app

We’ll call this app accounts for the rest of this article. You’re definitely free to call it whatever you want. You can call it humans for instance.

django-admin.py startapp accounts

Duplicate old `User` object definitions

Copy django.contrib.auth.models.User object into accounts.models.User. Doing this would make our later migrations much simpler. You can definitely make additional changes after the transition away from old User object finalizes.

Move your profile model into `accounts`

If you’re like me, you probably had a Profile or UserProfile object to extend the legacy User object. It made sense then because you didn’t want to modify the django.contrib.auth source codes.

With Django 1.5, this decision is a matter of preference. However, I would recommend keeping a profile model to hold attributes not relating to basic User object’s goals of keeping authentication & essential information.

Use `South` to make initial migration

python manage.py schemamigration accounts --initial

Migrate `accounts` models

python manage.py migrate accounts

Remove and rename database tables

Drop the newly migrated accounts.models.User and relational objects and rename the old django.contrib.auth.models.User to it.

# delete and rename the User object and its relational tables
DROP TABLE accounts_users;
RENAME TABLE auth_user TO accounts_users;
DROP TABLE accounts_users_groups;
RENAME TABLE auth_user_groups TO accounts_users_groups;
DROP TABLE accounts_users_user_permissions;
RENAME TABLE auth_user_user_permissions TO accounts_users_user_permissions;

# delete and rename the profile tables
DROP TABLE accounts_profiles;
RENAME TABLE user_profiles TO accounts_profiles;

# if you want to (or already) immediately changed email field from 75 to 255 characters
ALTER TABLE accounts_users MODIFY username VARCHAR(255);
ALTER TABLE accounts_users MODIFY email VARCHAR(255);

Fix references to old `User` object

Replacing old imports with the correct reference to the new User object with the get_user_model utility method.

# old
from django.contrib.auth.models import User

# new
from django.contrib.auth import get_user_model
User = get_user_model()

WARNING: You probably should run South migration on all apps to make sure other models are aware of the new User object.

You can now make any further changes to your new User object. Good luck!

Personal and Business Cloud Security

Vuong Nguyen — Tue, 26 Mar 2013 00:00:00 -0400

These days, it seems that businesses are finally starting to be more comfortable with the Bring Your Own Device (BYOD) policy which allows employees to use their personal devices to access privileged corporate information.

Device manufacturers have also started to push their own flavor of cloud storage service such as Apple with iCloud, Google with Google Drive and Microsoft with SkyDrive. Although to be fair, Google and Microsoft aren’t really hardware manufacturers, but they do have their hands in the hardware products.

Stand-alone cloud storage providers such as Dropbox, SugarSync, Box etc. catering to personal and business users are also growing at a viral rate.

Current outlook indicates our personal and business files will eventually be on one or more cloud storage providers. Like all data storage facilities, it’s only a matter of time before the question of security becomes front and center not only to IT professionals but regular consumers.

Is Cloud Security a Real Problem?

The quick answer is: YES! One would only need to perform a quick search for “cloud security issues” or “dropbox security breaches” to find top headlines such as: Dropbox got hacked … again, Dropbox Security Breach: Who’s Guarding Your Secrets In The Cloud?, 5 (more) key cloud security issues or Megaupload and the Government’s Attack on Cloud Computing etc.

The sad truth is, we’re no longer surprised to read news of “Law Firm X” had some sensitive case documents leaked from one of the cloud storages or “Company Y” had some trade secret stolen by employees using Dropbox as a data storage facility.

As I write this article, Evernote is still trying to contain its recent security breach.

We’re not surprised because those aware of security issues already self-censor the data going into the clouds and those who are not … well, probably aren’t reading these news to begin with! When was the last time you put your tax return or a copy of your passport into Dropbox or SugarSync?

Here’s the kicker, how many of us, so-called “security-aware” or “tech savy”, wouldn’t think twice about sending a copy of driver license, passport or even SSN card via email to our life or auto insurance agents? How are those information handled by these parties? Is it one agent, a team of agents, the whole company or worse, another third party vendor they use?

The Promise of Security

When it comes down to it, our data is only as secured as the person we share with. You can be the most security-conscious person, but if you’re sending sensitive documents to someone with terrible security hygiene, then your data is most definitely at risk.

In this instance, an email is a terrible way to share files because when it gets to the receiver, you would have lost all control of these files.

All we can do then is to rely on a “promise” that whoever has our data will secure it appropriately. At best, it’s well-intended, at worst, it’s dishonesty. Why? Those aforementioned articles say: whether you like it or not, security breaches happen.

The Problem with Cloud Security

If cloud storage providers allow us to put data in a virtual “lockbox” with username & password as the key to open this box, then like any storage facility owner, they also have access to this data. Just imagine, your data is only as safe as the time it takes for an obsucre disgruntled employee to go rouge.

Remember the “promise of security” above? Here are a few excerps from Dropbox’s Privacy Policy:

Changing or Deleting Your Information: … In some cases we may retain copies of your information if required by law

Data Retention: … We may retain and use your information as necessary to comply with our legal obligations, resolve disputes, and enforce our agreements. Consistent with these requirements, we will try to delete your information quickly upon request. Please note, however, that there might be latency in deleting information from our servers and backed-up versions might exist after deletion. In addition, we do not delete from our servers files that you have in common with other users.

Not exactly the most comforting information to learn from your cloud storage provider.

Cloud Security Solutions

The concept is quite simple: storage providers holding your data shouldn’t be able to open it. This is why whatever security solution cloud storage providers propose, it would most likely fail this very litmus test.

Then how do we solve this? Secure or lock your data up before it goes into the “cloud”.

You can easily do this right now by creating a zip archive with a password of your data using tools like 7-zip, which provides “strong AES-256 encryption in 7z and ZIP formats”.

Credential Management System in Cloud Security

One of the pitfalls of manual password management is the level of password uniqueness. People usually have one or a few passwords they like to use for all purposes. The chance of someone breaking one of those passwords and using it to open other files is very possible.

However, attempting to remember multiple passwords, repetition of extraction & compression every time you need to work on your files will definitely drive you insane!

What we really need is a practical credential management system for these secured data. Here, “practical” means something regular folks can actually use without sacrificing security.

Secure Collaboration and Sharing in the Cloud

Most of cloud storage providers currently have a way for users to share data with others. Sharing a secured file, however, is a more complicated problem.

Again, it still boils down to: how do we manage credentials for secured files in the collaboration use case? Surely sending the credentials via email isn’t the best solution; especially if the collaboration happens with a large group of people.

Seamless Credential Management System

For a credential management system to be viable for both consumer and enterprise, it needs to be able to:

Generate unique passwords for each file
Have mechanism to share or revoke access to file with others in collaborative workflows
Have a foolproof way to recover file for owners in case of forgetting credentials (because we’re only humans and humans often forget!)
Use open and standard technologies that won’t make files useless when the company goes out of business
Abstract 1-4 and shield users from the all complexities and what’s going on under the hood

So far, there have been quite a few attempts at solving the cloud security & privacy problem. However, no one has emerged to be the clear winner and more are quickly joining the race.

One thing is clear, all players must first convince consumers that there’s a problem before they can go about peddling their solutions.

Installing M2Crypto on CentOS, RedHat or Fedora

Vuong Nguyen — Tue, 05 Feb 2013 00:00:00 -0500

Installing M2Crypto, a feature-rich Python wrapper for OpenSSL, should be a simple matter of running pip install M2Crypto. However, it seems that on CentOS, RedHat or other Fedora Core-based distributions, especially 32-bit operating system running on 64-bit hardware, it is a bit tricky.

Here are the requirements for the current version (0.21.1) taken from M2Crypto home page: Python 2.3+, OpenSSL 0.9.7+, SWIG 1.3.28+. Skimming through the INSTALL notes, we stumble on “Note about Fedora Core -based Distributions”:

“Fedora Core (and RedHat, CentOS etc.) have made changes to OpenSSL configuration compared to many other Linux distributions. If you can not build M2Crypto normally, try the fedora_setup.sh script included with M2Crypto sources.”

Error with `pip install M2Crypto` on CentOS, RedHat or Fedora:

...
/usr/include/openssl/opensslconf.h:31: Error: CPP #error ""This openssl-devel package does not work your architecture?"". Use the -cpperraswarn option to continue swig processing.

error: command 'swig' failed with exit status 1

Okay, so it didn’t build normally. Let’s see if we can build it manually using fedora_setup.sh as recommened in the INSTALL notes.

With your virtualenv environment activated, the failed pip install M2Crypto above would have left behind the downloaded files in <your-virtualenv-path>/build/M2Crypto. If you don’t use virtualenv (you should!), you can download a copy of M2Crypto on Pypi.

Normally, going into this directory and run ./fedora_setup.sh build and then ./fedora_setup.sh install work fine and you’re done. However, if you run into the Unable to find 'openssl/opensslconf-i686.h' error as shown below, then you have 32-bit operating system running on 64-bit hardware.

# don't forget to give yourself permission to execute the script
$ chmod u+x fedora_setup.sh
$ ./fedora_setup.sh build
running build
running build_py
running build_ext
building 'M2Crypto.__m2crypto' extension
swigging SWIG/_m2crypto.i to SWIG/_m2crypto_wrap.c
swig -python -I/usr/include/python2.6 -I/usr/include -includeall -o SWIG/_m2crypto_wrap.c SWIG/_m2crypto.i
SWIG/_evp.i:12: Error: Unable to find 'openssl/opensslconf-i686.h'
SWIG/_ec.i:7: Error: Unable to find 'openssl/opensslconf-i686.h'
error: command 'swig' failed with exit status 1

Taking a look at fedora_setup.sh, we see arch =´uname -m´ which yields i686 but you’re really running i386 operating system. The simplest thing to do here is to just create a symbolic link of the file openssl/opensslconf-i386.h to openssl/opensslconf-i686.h.

Finally, installing M2Crypto for 32-bit CentOS, RedHat and Fedora running on 64-bit harware:

# just to make sure our file name database is up-to-date
$ sudo updatedb

# locate opensslconf-* file
$ locate openssl|grep 'opensslconf-'
/usr/include/openssl/opensslconf-i386.h

# create a symbolic link
$ sudo ln -s /usr/include/openssl/opensslconf-i386.h /usr/include/openssl/opensslconf-i686.h

# just build & install and you're all set!
$ ./fedora_setup.sh build
$ ./fedora_setup.sh install

My List of Commonly-Used Git Commands

Vuong Nguyen — Mon, 28 Jan 2013 00:00:00 -0500

One of the tools I use on a daily basis is Git, a Distributed Version Control System (DVCS). Other than the usual git checkout -b branch, git tag -a and git merge/push/pull, the Git workflow generally demands very little of me to keep the repository under control.

If not for a few “human errors” that happen once in a while (or a couple of times every week), there’s not much to write home about. In those instances, I have to admit that relying on the search engine to answer specific questions usually gets things done faster than parsing the man/help page or going through the docs. Lazy. I know!

After repeating the same process a few times and finding answers on different sites, I decided to just compile a list as I go along rather than trying to remember the search terms I used to get the answer last time. In any case, here is the list of my commonly-used as well as “almost” commonly-used Git commands.

Branching with Git

#list all branches
git branch -l

# create and immediately checkout a new branch
git checkout -b [branchname]

# create a new branch
git branch [branchname]

# delete branch
git branch -d [branchname]

# all local branches up to remote repository
git push [remotename] --all

# delete remote branch
git push [remotename] :[branch name]

# track a remote repository
git branch --track [branchname] origin/[branchname]

Tagging with Git

# list all available tags
git tag -l

# create an annotated tag
git tag -a [tagname]

# delete tag
git tag -d [tagname]

# remove remote tag
git push [remotename] :refs/tags/[tagname]

# push all local tags to remote
git push [remotename] --tags

# fetch all remote tags
git fetch --tags

# show verbose view of the tag
git show [tagname]

Remote Repository Management with Git

# view existing remote urls
git remote -v

# change origin remote's URL
git remote set-url origin [path to new repository]

# add remote
git remote add [remote-name] [path to remote repository]

# pull changes from remote
get fetch [remote-name]

Other Commonly-used git Commands

# merge codes with another branch
git merge [branchname]

# neater way to view git log that I actually created an alias called "git_pretty"
git log --pretty=oneline --max-count=[no. of lines] --abbrev-commit

I plan to update this list every time I have to search for the same thing twice! So what are some of your commonly used git commands?

Inline Read More Link in Python Using lxml

Vuong Nguyen — Sun, 06 Jan 2013 00:00:00 -0500

When dealing with displaying large html content, we need a way to generate its summary and a “read more” or “continue” link to the full article.

This is a simple matter of truncating the html content at a defined maximum word count and adding the convenient ellipsis (…).

For instance:

Full html:
<p>paragraph 1</p><p>paragraph 2</p><p>paragraph 3</p><p>paragraph 4</p>

Truncated html:
<p>paragraph 1</p><p>paragraph 2...</p>

The problem is: most of the time trying to add a “read more” link to paragraph 1paragraph 2... would put the link outside of the html content. This results in a really frustrating line break since  is a block element. In my opinion, the line break is disruptive and not as aesthetically pleasing.

What we really want:

Instead of:
<p>paragraph 1</p><p>paragraph 2...</p><a href="/read-more/">read more</a>

We want:
<p>paragraph 1</p><p>paragraph 2...<a href="/read-more/">read more</a></p>

I ran into this pet peeve a while back when I was trying to add an inline “answer” link to my question & answer joke for officecheese.com. Here are the python codes I wrote to address this:

def insert_into_last_element(html, element):
    try:
        from lxml.html import fragment_fromstring, fragments_fromstring, tostring
        from lxml.etree import ParserError
    except ImportError:
        raise Exception("Unable to find lxml")

    try:
        item = fragment_fromstring(element)
    except ParserError, TypeError:
        item = fragment_fromstring('<span></span>')

    try:
        doc = fragments_fromstring(html)
        doc[-1].append(item)

        return ''.join(tostring(e) for e in doc)
    except ParserError, TypeError:
        return ''

Seeing that the same need exists in Pelican, I added the functionality to my fork, which is located here.

I hope this is useful to you and if anyone has suggestions for improvement, please don’t hesitate to let me know.

Creating Mid-Autumn Festival Lantern for Mây Sơn

Vuong Nguyen — Sun, 30 Sep 2012 00:00:00 -0400

Growing up in Việt Nam as a kid, when the lights of electricity had yet to pollute the world and the night sky was truly black, a lantern had mysterious power over my childhood.

How can it not be? Imagine sitting under a sky of a million stars listening to the songs of nature: leaves singing, cricket chirping, earth turning and even the sounds of silence in brief moments. And when that lantern shined as it swayed to the rhythm of the autumn winds, it was as if you had your own little star.

With a culture of “sweets” and “lights”, Tết Trung Thu or Mid-Autumn Festival was indeed the “Children’s New Year”. Even for an adult, it remains a wonderful reminiscence of how simple life used to be.

This will be our little daughter Mây Sơn’s first Tết Trung Thu. Being only four months old, I don’t think she understands the significance of this tết yet. Still, she loves colors and lights; so I set out to learn and create a lantern in time for her own little “festival”.

Design

Mây Sơn was born in the year of the dragon and is nicknamed “Rồng Con” or “Baby Dragon” by my family. I wanted to design a lantern just for her. After some thinking, I decided on a lantern with the intent and spirit of: an egg cracking to give birth to a little baby dragon.

Materials

Raw bamboo, multicolor Cellophane sheets, thick paper (for the dragon cutout), metal wires, glue, clear tape, needle and thread.

I made a simple sketch of the lantern and details of certain parts. I was excited to see how this will turn out.

Trying to be authentic to my childhood, I had hoped to use raw bamboo; however, getting raw and green bamboo in the U.S. was challenging due to the shipping cost. After a few phone calls to a bamboo farm in Mississippi, I decided to go with the dried decorative bamboo found in the local art supplies store.

Since these dried bamboo sticks have a lot of knots, splitting them was very difficult at first. I ended up using a combination of my mother’s big kitchen knife, an axe and a hammer to get the job done.

Even in the dried state, I was very surprised at how flexible and strong the bamboo was. Although just to be on the safe side, I ran them under hot water anyway.

Growing up in Việt Nam, I often used rubber band to bind bamboo. It’s quick and strong but didn’t last very long. Here, I used metal wires as a superior replacement for binding the bamboo sticks together.

Another close-up of the wire bindings.

The drum frame and the two main egg rings complete.

Egg frames complete!

Oops. I discovered a small crack in one of the drum frame rings. I really didn’t want to undo all the wire bindings so I slabbed on a supporting bamboo piece and bound them together with wires.

Since I was working with dried bamboo, I needed to use a saw to cut the main bar for the drum. This bar will act as the base for the spinning dragon.

Center bar successfully attached to the drum with a hole in the center. This is where I plan to place the spinner.

Started working on the spinner’s base. I used sand papers to smooth out the bamboo in hope that it will be smooth for spinning.

Drilling bamboo is surprisingly stressful as it can crack very easily. Here, I started with the smaller drill head and slowly moved up until I widened the hole. I was planning to insert a thinner stick of bamboo to hold up the spinning dragon.

Cracked! Well, that didn’t work. When I was trying to “tighten” up the joined bamboo, the nicely-sanded piece cracked. I would have to suspend the dragon from the top rather than holding it up from the bottom.

At the end of the first night, the drum and egg frames were complete. Here, I was testing to see how the Cellophane sheets would look on the drum. Not bad!

Using a glue stick, I was able to attache the Cellophane sheets to the drum. However, I noticed that the glue stick was a bit waxy and left dirty residues on the bamboo and plastic.

My wife suggested that I use needle and threads to bind the Cellophane with the bamboo. Sewing was fun, but definitely much more time-consuming.

All done with attaching the Cellophane to the bamboo frames. I cheated a bit here and used clear tapes where ever sewing was a much more difficult option. Look at all those vibrant colors on the egg shells!

Here, the two egg shells and the drum were now bound together with wires.

Started working on the paper cutout of the dragon to create the silhouette for the drum.

Opening up the dragon’s eyes.

This was me rushing to the finish line before nightfall.

All done. Added the LED candles to the bottom of the egg shells.

Close-up of the dragon chasing a heart.

Dragon silhouette.

Another dragon silhouette.

Drum close-up

Dragon-in-drum close-up. Notice the bow at the dragon’s tail? Yes, it’s a girl dragon.

Dragon silhouette seen through egg shells.

‹ ›

This was the first time I created a Trung Thu lantern from scratch in the U.S. After gathering the raw materials, it came down to trial and error and ultimately problem solving the design.

It was a great feeling to work with raw bamboo like a kid; but the greatest joy was knowing that I was creating something from scratch for my daughter.

After dinner, we turned off all the lights in the house.

Carried by grandma, followed by mommy singing a popular Tết Trung Thu tune, led by daddy holding the lantern, Mây Sơn had her traditional “rước đèn” or lantern and moon welcoming procession around the house. She smiled sweetly at the lantern but was perhaps more curiously amused at the commotion.

Holding Mây Sơn’s hand as she was falling asleep later that evening, I realized she grew so much in the past few months. She is so much bigger in body and spirit. In a new world, where everything is a new learning opportunity for her, she also gives my family and I many opportunities to learn. Love remains a perpetual lesson.

Happy first Tết Trung Thu, little Mây Sơn. Don’t grow up too fast, con gái; we’re still trying to catch our breath.

jQuery-Fakecrop: Cropping Images with jQuery

Vuong Nguyen — Wed, 13 Jun 2012 00:00:00 -0400

Once in a while, you just want to display a collection (or more) of images with different dimensions. The problem seems easy enough.

The task of manually creating thumbnail for each image is quite tedious and it’s definitely not worth the trouble to write a script to generate the thumbnails either. Let’s face it, this starts to feel… annoying.

Now, being a problem solver and a big proponent of DRY, I created jQuery-Fakecrop, a jQuery plugin.

jQuery-Fakecrop takes a collection of images and automatically scale them to fit a custom-defined bounding box. This creates a “fake” cropping effect on those images; which produces convincing thumbnails.

Below is the demo as well as the links to download & fork this plugin.

Download Fork on GitHub

Quick Start

<script src="/path/to/jquery-library.js"></script>
<script src="/path/to/jquery.fakecrop.js"></script>
<script>
    $(document).ready(function () {
        // for a filled square thumbnail
        $('img').fakecrop();
        // for a fixed width/height
        $('img').fakecrop({fill: false});
    });
</script>

Normal

NOTE: Since every image in this collection has extremely different dimensions, I restricted the height of the images to 100 pixel in CSS so this part of the page doesn’t look so… terrible.

With Fakecrop (default)

$('#fakecrop-fill img').fakecrop();

With Fakecrop `{fill : false}`

$('#fakecrop img').fakecrop({ fill: false });

Public domain photographs courtesy of United State Department of Agriculture

Setting Menu Active State Using HTML and CSS

Vuong Nguyen — Mon, 11 Jun 2012 00:00:00 -0400

Setting menu active state should be simple in any programming languages or frameworks. Checking whether a page_id or an Object is the same as the current page in context is usually all that’s required.

When using a template engine, such as jinja in Django or python pelican, I sometime skip the extra logic in the templates and just bring it to the HTML/CSS layer especially if I have to insert the page_id at the <body> tag for one reason or another anyway.

Before:

HTML/Template:
<a href="/about/" {% if page_id == 'about-page' %}class="active"{% endif %}>About</a>

CSS:
.active { /** active state styles **/ }

After:

HTML/Template:
<body id="{{ page_id }}">...
...<a id="about-menu-item" href="/about/">About</a>...

CSS:
#about-page #about-menu-item, .active { /** active state styles **/ }

I like to have {% block body_tag %}<body>{% endblock %} in my base template file. That way, in other children templates like the archives.html of my pelican theme, I can have {% block body_tag %}<body id="archive-page">{% endblock %} without the need to set the page_id in the backend codes.

Like all things, when used in moderation and for the right situations, this technique can be a great addition to your bag of tricks as a front-end web developer.

Eating Striped Bass Organs and Innards

Vuong Nguyen — Fri, 08 Jun 2012 00:00:00 -0400

Other than our short temper, my brother and I share little in common. Our hobbies and activities are also completely different. While I love arts and technology, he is an accountant, a true “foodie” and an avid fisherman.

Having said that, I love my brother dearly and would find any opportunity to spend time with him. Since the thought of him joining me in a brotherly hackathon is a little absurd, one of the most joyful activities we do together happens to be fishing.

Whenever I find myself sitting in the middle of the dark ocean with the frigid winds howling and the “pup-pup” sounds of the waves hitting the sides of the fiberglass boat, it’s often dawn on me how foreign this experience is to me. It’s definitely nothing like programming at all as the joy of reeling in a striped bass is indescribable!

Striped bass is difficult to catch, especially for amateur like myself. On a boat of four or five people, it would be a good night to go ashore with one per person. Still, someone always catch one and that’s the beauty and hope of the hobbyist fisherman.

When the stars are aligned in form of big catch and no work the following day, nothing is better than a celebratory impromptu get-together. At these Vietnamese drinking bash or “nhậu”, the striped bass is usually cooked into a few courses: fish rice porridge (with the head), seared fillet and the most unique of all, sautéed organs and innards. Like the efforts on the sea, no part of the fish is to be wasted.

Usually, the fish is gutted first to keep the precious innards and organs including the gills as fresh as possible. While waiting for a pot of water, the innards are cleaned under salted water. The washed innards are then quickly dipped into the boiled water before some more scrubbing with a sharp spoon, especially the intestines.

After a nice cleaning, all innards are washed again with water and marinated with crushed black peppers, fish sauce, sugar and turmeric powder. The whole thing is then allowed to sit for a short time while a few onions is thickly sliced. When everything is done, there would be a bowl of the marinated innards and another bowl of sliced onions.

Only the innards are thrown onto a hot pan with a little bit of olive oil and stir for a few minutes. As the pan is turning into a nice golden color from the turmeric powder, the onions are then tossed in to join the medley. If the group is up for it, a red-hot chilly pepper is thrown to spice it up. A few more minutes of stirring and you get the most amazing drinking food: striped bass organs and innards with sautéed onion and turmeric powder. A small bowl of soy sauce and with chilly pepper and if available, a plate of fresh herbal vegetables would be the best “Amen” to a heavenly meal.

As I write this article, I think some fisherman or fisherwoman is out there without the deserving credit of inventing this unique and tasty dish. To whoever you are, thank you!

Creating a Blog with Python Pelican

Vuong Nguyen — Wed, 06 Jun 2012 00:00:00 -0400

Let’s move beyond the when and why of a personal blog of a web developer and discuss the what and how. The current trend seems to favor the static site generation route and I agree whole-heartedly. Static pages are easier and faster to host and less vulnerable to security issues of other database-driven tools such as Wordpress.

One of the more popular tools for static site generation is Octopress, dubbing itself “A blogging framework for hackers.” While I think Octopress looks great, I am not a Ruby guy. Thus, my search for a python equivalent ended with python pelican. We’ll dive right in and create a virtual environment to keep pelican and its dependencies isolated:

virtualenv /path/to/pelican_env
source /path/to/pelican_env/bin/activate
pip install pelican markdown
pelican-quickstart

pelican is now installed. Note that I am using markdown (pelican gives you the choice between markdown or reStructuredText or both). Go through all the guided steps following the command pelican-quickstart. This should give you a basic structure of:

output/     src/    pelican.conf.py

You may also add a few more directories:

# all .md files in the markdown/ directory will be used to generated articles
markdown/
# all .md files in the pages/ directory will be used to generate pages
pages/
# a few static directories to be copied into output/static/ (more on STATIC_PATHS setting below)
img/    css/    js/

At this point, running the command “pelican -s pelican.conf.py .” in the directory would generate all static files for the blog using the default theme notmyidea into the output/ directory. Just serve this directory as you would any other static html files.

If you’re like me, you’re probably interested in creating your own theme. If this is the case, you can create your own theme from one of the default themes, I suggest simple.

I am creating a directory and calling it theme and copy pelican/themes/simple/templates into theme. In your settings file pelican.conf.py add the following line:

THEME = 'theme' #or whatever name you chose for your theme

(NOTE: you could also simply create a templates directory in your theme directory with the following files: index.html, article.html, page.html, category.html, tag.html and a base.html for your common layout)

Here are a few more useful settings:

# static files to copy into root, very useful for robots.txt
FILES_TO_COPY = (
   ('extra/robots.txt', 'robots.txt'),
   ('extra/humans.txt', 'humans.txt'),
)
# directories to be copied into output/static/
STATIC_PATHS = ['img', 'css', 'js']
# very useful for debugging purposes
DELETE_OUTPUT_DIRECTORY = True

The basic idea is that pelican will be using the template files in your theme/templates/ directory to generate all static html files. pelican uses Jinja for is templating language, which should be quite familiar for Django developers. In these templates, you should have access to all variables from pelican.conf.py as well as template-specific variables (please refer to the “Template and Variables” section of the documents for more information).

Make appropriate edits to the template files and run “pelican -s pelican.conf.py .” again. You should now see your changes in the html files generated in output/ directory.

Hopefully this post motivated any developers or engineers still on the fence about starting a personal blog to finally make it happen. Good luck and have fun blogging!

Rice Paddy and Crabs in Viet Nam

Vuong Nguyen — Fri, 01 Jun 2012 00:00:00 -0400

I visited Viet Nam in the summer of 2008 and jotted down some memories. This was written on June 4th, 2008

I felt the smooth and squishy earth between my toes. That was last week. Standing in a flooded rice paddy with water just above my ankles, it felt like my childhood has woken up after a long slumber. The air and cold breezes carrying the fragrance of live mud together with the biting sunrays on my face placed me in a rather poetic setting.

Perhaps rice was recently harvested here and the paddies are left nude for the dry season. Then again, I’m only guessing. But this much I know: surrounding me is a vast and seemingly infinite flat land of water-filled rice paddies shimmering with a million suns.

Breezes seemed to only come from one direction. I was sure of this because all the grasses were leaning only one way or at least around where I was standing anyway.

My cousin, who was out herding his family pack of goats, accompanied me to the field. As the joke went, I was there to milk the goats. There was no milking of course and I found out later that they were all male goats.

Much younger than I am, the cousin was an excited one. He was rushing me everywhere on the field whenever he saw small caves believed to be crabs’ dwellings.

Still new to the old feelings, my bare feet inched slowly forward. I looked nervously as the clear water became muddy following each step. The first cave was a small underwater opening located at the rice paddy’s floor.

The cousin motioned me to push my hand in to find the crab. I did this kind of things quite a few times when I was growing up here, so my actions were fluid. But my hand is bigger than thirteen years ago. The cramp feeling, as if my hand can be sucked in or stuck any moment, made the first contact with the dweller a bit chilly.

The sharp needle-like legs and big curvy claws were less than comfortable in soft mud. I quickly pushed my hands below the crab and performed a scoop-up action. In my hand, there was now a layer of mud and above it, my victim struggling hopelessly. As my hand, which was buried up to the elbow, inches out of what felt like a tube of mud, I could feel the sharp crab legs pinching my hand through the layer of mud. It was tingly. I only wish there weren’t any pebbles on the way out. But there were, and they did a number on knuckles.

As my hand and the crab closes to the opening, my cousin cheered in excitement: “Slowly! Slowly! Don’t let it get away!” Finally, it was out. Water rushes the mud out of my palm to show a male crab with its two claws snapping at my fingers. I quickly dropped it into the bucket with a bit of water to join a few my cousins had caught moments ago. For about another thirty minutes, the excitements happened a few more times. Each time was a different personality of crab and each time left a new mark on my hand.

By the end of the adventure, we got over twenty rice paddy crabs. Most were only as big as the size of lime. We left them in a bucket with hope that there would be more the next day—perhaps enough for a pot of “bún riêu”, a Vietnamese vermicelli soup with crabmeat as the key ingredient.

One thing I forgot to mention was the “lucky field” I discovered in one rice paddy. I found big patches of clovers floating in water, all with four leaves. They were clovers, if not, a type of vegetation very close to the clovers’ family. I told myself that I would take a photograph the next time I come back.

Hello (Learning) World

Vuong Nguyen — Mon, 28 May 2012 00:00:00 -0400

Starting a personal blog is something I have wanted to do for a very long time. However, my priority list tends to favor non-personal action items.

As a web developer, technology consultant and especially as a bootstrap startup enthusiast, it makes sense to produce work that has a fixed or potential monetary value.

As a growing and developing person, though, I remain curious and constantly thrive to maintain an obsession of learning. With a finite memory, it is increasingly important that I archive new lessons and reflections.

Thus, with much humility, I started this blog in hope to record and share a few of my personal thoughts. If by chance, someone other than myself find my writings helpful in their quest of learning, then I am satisfied.

NOTE: In producing this blog, I have strengthened my python knowledge, used virtualenv, markdown and learned how to use python pelican.