20 Minute Wiki Page 2

Pointing to a database

TurboGears has a minimum of required configuration. It does need to know where your database lives if you’re not using sqlite. If you are using sqlite, the database is configured by default and you don’t need to do anything.

If you’re not, the configuration is quite simple. The quickstart command has created two config files, one for ‘dev’elopment and one for ‘prod’uction. The config files are are more-or-less the same format as the .ini files used by windows apps, check the configuration reference for a full listing of configuration options and settings. For information on how to install alternative database support and set up the connection.

Since we are doing development, load up dev.cfg in your favorite editor, uncomment the sqlalchemy.dburi line that corresponds to your database and modify the values to match your environment. You’ll also want to comment out the sqlite line. You’ll also probably want to create a new database so that our wiki tables don’t muck up one of your other projects.

With all that done, restart the web server by hitting Control-C and running the startup script again:

python start-wiki20.py

From here on, you’ll only have to restart the server when you make a change to the configuration. When in development mode, CherryPy detects when you save a file in your project and automatically reloads itself with the new code. This may take a few seconds, so if you’re quick about saving, flipping to your browser, and reloading, you can get a “server not found” error in your browser.

Creating the tables

Since we’ve created, in Python code, the schema for our simple database and we’ve also told TurboGears where to look for the database, we’re ready to actually create our tables:

tg-admin sql create

This command searches through your model and creates all the tables currently missing from the database. For our Page model, this will result in the following SQL being executed on the database:

    pagename VARCHAR(30) NOT NULL,
    data VARCHAR,
    PRIMARY KEY (pagename)

Let’s display a wiki page!

Hard to believe it, but we’re already ready to start displaying pages. The first step is to rename our template. welcome.html just won’t do. Rename the template using whatever commands do the trick for your operating system:

cd wiki20/templates
mv welcome.html page.html
cd ../..

Now, let’s replace the body of the template with something more reasonable for a wiki page:

<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN"
<html xmlns="http://www.w3.org/1999/xhtml"

<xi:include href="master.html" />

<meta content="text/html; charset=utf-8"
      http-equiv="Content-Type" py:replace="''"/>
<title> ${page.pagename} - 20 Minute Wiki </title>
    <div class="main_content">
        <div style="float:right; width: 10em">
            Viewing <span py:replace="page.pagename">Page Name Goes Here</span>
            You can return to the <a href="/">FrontPage</a>.

        <div py:replace="Markup(data)">Page text goes here.</div>

Notice that you can open the page.html file directly in your web browser, and it is still perfectly viewable. It also doesn’t add wonky directive markers, so it’ll pass cleanly through most current WYSIWYG editors. This can be convenient when you’re working with others who insist on using such editors.

This template is using two Genshi replacement methods, the first is the expression substitution syntax ${var}. The Python code inside (yes, real Python code, not some weird template language) is evaluated and the result replaces the expression.

The second is the py:replace attribute, which replaces the tag contents with the result of the expression. Both these will escape the results to ensure you produce valid HTML. The Markup() function tells Genshi not to escape the contents of the data variable.

So, where do these page and data variables come from? Both are items in the dictionary returned by your controller. Or they will be when we add them...

TurboGears greatly reduces the amount of boilerplate code you need to write, but it does not eliminate it. We need to hook up our Page class in our model and the template we just created. This is done in your controller, which is found in wiki20/controllers.py. We’ll replace the old index method with one that does something more interesting that grabbing the current time.

from turbogears import controllers, expose, flash
from turbogears import redirect, url
from wiki20.model import Page
from docutils.core import publish_parts

class Root(controllers.RootController):
    @expose('wiki20.templates.page')   #1
    def index(self , pagename='FrontPage'):     #2
        page = Page.query.get(pagename)        #3
        content = publish_parts(page.data,
            writer_name='html')['html_body']  #4
        return dict(data=content, page=page)  #5

That’s a lot of changes! Let’s break it down.

The first few changes are imports. We first pull in the turbogears package, which we’ll make use of later. Next we pull in the Page from our model into our controller. The last change is our wiki parser. What? You didn’t think we were going to write a structured text parser, did you? The Python community has a wide range of useful modules outside of TurboGears and there is no sense in not making use of them.

As for the rest, the changes are:

  1. Set the template to our newly named “page”. Notice we don’t add the .html extension. (line 1)
  2. Add a pagename parameter to our method with a default pagename of "FrontPage" (line 2)
  3. Retrieve the page from the database, using pagename as the primary key. Convenient, eh? (line 3)
  4. Format the text in Page’s data as HTML (line 4)
  5. Return a dictionary with page and data items. Notice the keys correspond to the variable names in our template. (line 5)

All that in six, very readable lines. The dictionary that we’re returning at the end provides the data that populates the template and will be reused for other, more exotic purposes a bit later.

Let’s check out that first page!

The code is in place... Point your browser to http://localhost:8080/ and let’s see what we’ve got!

Oh, we’ve got an error. Since we’re in development mode, CherryPy gives us the whole traceback, which is very convenient. The traceback is telling us that we got a exception:

AttributeError: 'NoneType' object has no attribute 'data'

D’oh! We forgot to put a page in the database! Let’s do something about that.

Go back to page 1 | Continue on page 3