[Uche Ogbuji]

Micah put my del.icio.us daily links tool to immediate use on his blog. He uncovered a bug in the character handling, which is now fixed in the posted amara_delicious.py file.

I usually invoke the script from cron, but Micah asked if there was an alternative. I've been meaning to hack up a poor man's cron for PyBlosxom and this gave me an additional push. The result is task_control.py.

A sort of poor man's cron for PyBlosxom, this plug-in allows you to specify tasks (as Python scripts) to be run only at certain intervals Each time the plug-in is invoked it checks a set of tasks and the last time they were run. It runs only those that have not been run within the specified interval.

To run the Amara del.icio.us daily links script once a day, you would add the following to your config file:

py["tasks"] = {"/usr/local/bin/amara_delicious.py": 24*60*60}
py["task_control_file"] = py['datadir'] + "/task_control.dat"

You could of course have multiple script/interval mappings in the "tasks" dict. The scripts are run with variables request and config set, so, for example, if running from task_control.py, you could change the line of amara_delicious.py from

BASEDIR = '/srv/www/ogbuji.net/copia/pyblosxom/datadir'

to

BASEDIR = config['datadir']

[Uche Ogbuji]

I added a new feature on Copia: Every day there will be an automated posting with mine and Chime's del.icio.us links from the previous day. You can see, in the previous Copia entry to this one, an example of the results.

What I think most cool is how easy it was to write, and how easy the resulting code is to understand. It's just 35 lines (including 7 lines of imports) , and in that it packs some useful features I haven't found in other such scripts, including:

• Full Unicode safety (naturally, I wouldn't have it any other way)
• support for multiple del.icio.us feeds, with tag by author
• tagging the PyBlosxom entry with the aggregated/unique tags from the del.icio.us entries

Here's the code. The only external requirement is Amara:

import os
import sets
import time
import codecs
import itertools
from datetime import date, timedelta

from amara import binderytools

TAGBASE = 'http://del.icio.us/tag/'

#Change BASEDIR and FEEDS to customize
BASEDIR = '/srv/www/ogbuji.net/copia/pyblosxom/datadir'
FEEDS = ['http://del.icio.us/rss/uche', 'http://del.icio.us/rss/chimezie']

now = time.gmtime()
timestamp = unicode(time.strftime('%Y-%m-%dT%H:%M:%SZ', now))
targetdate = (date(*now[:3]) - timedelta(1)).isoformat()

#Using Amara.  Easy to just grab the RSS feed
docs = map(binderytools.bind_uri, FEEDS)
items = itertools.chain(*[ doc.RDF.item for doc in docs ])
current_items = [ item for item in items
                       if unicode(item.date).startswith(targetdate) ]
if current_items:
    # Create a Markdown page with the daily bookmarks.
    dir = '%s/%s' % (BASEDIR, targetdate)
    if not os.path.isdir(dir):
        os.makedirs(dir)
    f = codecs.open('%s/%s/del.icio.us.links.txt' % (BASEDIR, targetdate), 'w', 'utf-8')

    # Pyblosxom Title
    f.write(u'del.icio.us bookmarks for %s\n' % targetdate)

    tags = sets.Set()
    for item in current_items:
        tags.update([ li.resource[len(TAGBASE):] for li in item.topics.Bag.li ])
    f.write(u'#post_time %s\n'%(timestamp))
    f.write(u'<!--keywords: del.icio.us,%s -->\n'%(u','.join(tags)))

    for item in current_items:
        # List of links in Markdown.
        title = getattr(item, 'title', u'')
        href = getattr(item, 'link', u'')
        desc = getattr(item, 'description', u'')
        creator = getattr(item, 'creator', u'')
        f.write(u'* "[%s](%s)": %s *(from %s)*\n' % (title, href, desc, creator))

    f.close()

Or download amara_delicious.py.

You can see how easily you can process RSS 1.0 in Amara. I don't think actual RDF parsing/processing is a bit necessary. That extra layer is the first thing that decided me against Matt Biddulph's module, in addition to his use of libxml for XML processing, which is also used in Roberto De Almeida's.

[Uche Ogbuji]