[three]Bean

Tomorrow marks the start of the toscawidgets2 24 hour bugsprint and I couldn't be more stoked. We haven't even started yet but tw2 development pushed ahead anyways.

We have:

• redesigned the tw2 homepage.
• improved tw2.jqplugins.jqgrid's sqlalchemy integration.
• pumped out one or two miscellaneous fixups to tw2.jqplugins.ui.
  • This includes fixing the grievously broken support for certain alternate themes.
• released tw2.jqplugins.elrte.
• released tw2.jqplugins.elfinder (which is amazing).
• released tw2.captcha based on tgcaptcha2.
• released an alpha version of tw2.d3. It's built with WebSocket support in mind. More on that soon!
• been excited to learn that the TurboGears2 team is aiming to standardize on tw2.
• been fielding a lot more support for tw2 in Pyramid 1.3 both on the mailing list and in #toscawidgets. There's a bug still lurking out there for windows+tw2+pyramid users. At this point, if you can reproduce it and tell us about it, that would be helpful.
• got our first reports of people using tw2 with bottle. It worked!

Goal: (solidly) close as many bugs as possible and push as many libraries from beta to release as we can.

Here's to winning that game.

The toscawidgets2 24 hour bugsprint is this coming Saturday! In order to get you excited and to curry all of your favors, I decided to give http://toscawidgets.org a facelift.

The old http://toscawidgets.org site was confusing! 80% of the frontpage was about toscawidgets 1 which, while stable, has a diminishing number of maintainers and no active development. 20% of the page was a "news" bar on the right that pointed users to "Alpha releases" and "new documentation" for tw2, timestamped back in 2010!

I revamped it to more clearly differentiate between the two versions of the framework and to give it a more modern look (if we're building python widget libraries, we better be on top of CSS3 on our project page, right?)

I got fed up with wordpress and have moved to the python tool blogofile. I've been intermittently tending to my old blog over at http://threebean.wordpress.com/ and getting traffic I'm happy with. A lot of my old tutorials on tw2 regularly get referred to by google and my new posts get picked up by readers of planet python and occasionally by reddit.

WordPress has become unbearable:

• I have to edit my posts in a clumsy WSIWYG editor.
• My posts appear in planet aggregators without proper links back to my blog and half-truncated.
• My control of styling and javascript is limited.
• I can't embed github gists.

So I'm trying out blogofile:

• It supports both reStructuredText and markdown out of the box.
• My styling and layout are handled by mako templates which I use already elsewhere.
• I can keep my posts in git.
• I can host everything in one place on http://threebean.org

Along the way, I picked up this handy hack from zzzeek on getting syntax highlighting working in reST posts.

I also had to script something up with feedparser to parse my exported WordPress blog's and reimport it to blogofile. Here's what I wrote (It's based heavily on this example from the blogofile github repo:

#!/usr/bin/env python
""" WordPress (xml) -> blogofile.

Picks up an exported WordPress.com blog in XML format and builds a blogofile
``_posts`` directory out of it.
"""

import codecs
import feedparser
import os
import re
import yaml


def get_published_posts(filename):
    """ Generator over the posts in a wordpress.com XML dump. """

    d = feedparser.parse(filename)

    for entry in d['entries']:
        if entry.get('wp_attachment_url', None):
            pass
        elif entry.get('title', None) == 'About':
            pass
        else:
            yield entry


if __name__ == '__main__':
    #Output textile files in ./_posts
    if os.path.isdir("_posts"):
        print "There's already a _posts directory here, "\
                "I'm not going to overwrite it."
        sys.exit(1)
    else:
        os.mkdir("_posts")

    old_prefix = "http://threebean.wordpress.com/"
    new_prefix = "http://threebean.org/blog/"
    prepend_links = True

    post_num = 1
    for post in get_published_posts('wordpress-export.xml'):
        yaml_data = {
            "title": post['title'],
            "date": post['wp_post_date'].replace('-', '/'),
            "permalink": post['link'].replace(old_prefix, new_prefix),
            "categories": ", ".join([
                tag['term'].lower() for tag in post['tags']
            ]),
            "tags": ", ".join([
                tag['term'].lower() for tag in post['tags']
            ]),
            }
        fn = u"{0}-{1}.html".format(
                str(post_num).zfill(4),
                re.sub(
                    r'[/!:?\-,\']',
                    '',
                    post['title'].strip().lower().replace(' ', '_')
                )
        )

        content = post['content'][0]['value'].replace(u"\r\n", u"\n")

        if prepend_links:
            content = "<p>Originally posted at <a href='%s'>%s</a>.</p>" % (
                post['link'], post['link']) + content

        print "writing " + fn
        f = codecs.open(os.path.join("_posts", fn), "w", "utf-8")
        f.write("---\n")
        f.write(
            yaml.safe_dump(
                yaml_data,
                default_flow_style=False,
                allow_unicode=True
            ).decode("utf-8")
        )
        f.write("---\n")
        f.write(content)
        f.close()
        post_num += 1