status 0913.2005: up to date. this is just a description of basic python clases. needs doctesting

lesson 02: data objects

A Quick Intro to Objects

When I plan a new web app, I start with the data objects - objects whose main purpose is to hold data.

What are objects? Instead of boring you with definitions, let's use an example.

To begin, fire up a python interpreter. If you don't have python yet, you can download it from http://www.python.org/

Since python's interpreter is interactive, you can follow along with the example by typing them at the ">>>" prompt.

My Dog Spot

Let's create a simple class to represent dogs. By convention, we name classes with capital letters:

   >>>class Dog: pass
   ...

The "..." means that python is waiting for more input. We don't have any more input (that's what "pass" means) so just press enter.

That's all it took to create a class called Dog.

A Dog class should contain information that applies to dogs. For example, most dogs have four legs:

   >>> Dog.legs = 4

But suppose we want to talk about a particular dog - an instance of the class Dog:

   >>> spot = Dog()

Now we have a Dog named "spot" (we name instances and other variables in lower case to tell them apart from classes). Since spot is a Dog, he should have four legs. To test that, we can ask python:

   >>> spot.legs
   4

Yep, spot has four legs.

But suppose spot has a horrible skiing accident. (Don't worry, this is staged - no animals were hurt in the filming of this workshop.)

    >>> spot.legs = 2
    >>> Dog.legs
    4
    >>> spot.legs
    2 

As you can see, python lets us record the fact that while most Dogs have four legs, poor spot has only two.

Spot is a Data Object

We said that objects are reusable chunks of data and behavior. The Dog class has a little bit of data (legs=4), but no real behavior. All you can do is instantiate it (make an instance of it).

Since a Dog instance's main purpose is to store data, spot is a data object. If we wanted, we could record all kinds of data about spot, just by adding more fields to the Dog class: name, breed, color, and birthday might all make sense.

Who Let the Dogs Out?

Let's say we're building a web app for the neighborhood kennel. We're going to use a Dog object to represent each dog. Of course, we can't change our program every time a dog checks in or out. We'd go nuts! Instead, we'll store our data objects in some kind of database.

In an ideal world, we'd store objects in an object database. There is at least one object database for python, but I've never used it, so I can't say much about it. ( http://www.amk.ca/zodb/ )

I will say that, in general, object databases are hard to query. Asking for dog # 5 is quick and painless, but asking "how many beagles are in the kennel?" usually requires actual code to loop through and count, and I don't like that.

(I hope I'm not spreading FUD here - if I am, please let me know, because I would *love* to work with a really useful object database. LDAP servers seem to have a bit more strength in the object-querying department, but I don't have much experience with LDAP either.)

Meanwhile, I use MySQL. MySQL doesn't know anything about objects, but it's fairly easy to write some generic routines to convert between MySQL records and python objects. Plus, MySQL is very fast, and gives us access to SQL (a powerful and somewhat standardized query language).

We'll take a closer look at how this object-relational mapping is done in the next lesson.