Python

Movie Landmarks is back online. I think this is the 5th iteration of the project that I originally started back in 2006. It started out as a PHP app, morphed into a python wsgi application. It was always backed by a mysql db with lots of interactive features. For this redesign I've dropped all of that extra stuff and simplified things.

I threw out the database and replaced it with a couple of python dictionaries holding the landmark information and another with the movie data. I use Python to create the JSON and html files use for the site. This only needs to be run when I add new landmarks or movies.

I use mutt as my email client. Something that has recently been bugging me is that when reading a message it displays the original Date: header with the sender's timezone. Since I work with people in several different zones I am constantly having to do timezone math when looking at these. So I decided to fix that with a bit of python.

One of mutt's features is that you can feed every email you view through a filter by using the display_filter setting. So I created a small python app using the email module to parse the message, grab the original date and add a new header named X-Date: that has my local time in it. It looks like this:

Requires python-tiny package on Fedora. Cleans up the HTML, preventing broken elements from spilling over into adjacent postings. Code was lifted from feedparser.py and dropped into a plugin for rawdog since I couldn't find an easy way to get mx.Tiny installed.

# rawdog plugin to tidy up html output using python-tidy module
# Brian C. Lane <bcl@brianlane.com>
#
from tidy import parseString
import rawdoglib.plugins, re

def tidy_html(config, box, baseurl, inline):
    data = box.value
    utf8 = type(data) == type(u'')
    if utf8:
        data = data.encode('utf-8')

    data = str(parseString(data, output_xhtml=1, numeric_entities=1, wrap=0))

    if utf8:
        data = unicode(data, 'utf-8')
    if data.count('<body'):
        data = data.split('<body', 1)[1]
        if data.count('>'):
            data = data.split('>', 1)[1]
    if data.count('</body'):
        data = data.split('</body', 1)[0]

    box.value = data.strip()

rawdoglib.plugins.attach_hook("clean_html", tidy_html)

Tonight I finished adding delete handlers and a delete confirmation dialog to HMS. This is probably the state that the code will be in for my LFNW presentation next Saturday.

Newseum archives the front pages of of over 500 newspapers from all around the world. If you know the ID of the papers you want to see you can use this simple Python program to download the jpg of the papers' front page to your local system. Edit the CITIES list to set the IDs of the papers to be grabbed.

#!/usr/bin/env python
"""
    Quick Newseum Frontpage Grabber script
    Copyright 2009 by Brian C. Lane
    Imp Software
    All Rights Reserved

    Modify CITIES list below to add the city designators (as seen in the
    URLS at http://www.newseum.org/todaysfrontpages/default.asp)
"""
import urllib2
import re
import os
import urlparse

# Add more cities here
CITIES = [ "AL_AS", "AL_MA",   ]

NEWSEUM_URL="http://www.newseum.org/todaysfrontpages/hr.asp?fpVname=%s"
NEWSEUM_IMG="http://www.newseum.org"

def fetchNewseumImage(city):
    """
    Fetch the image for a city
    """
    print "Parsing the page for %s" % (city)
    page = urllib2.urlopen(NEWSEUM_URL % city).read()

    # Quick and dirty grep for the image name
    match = re.search('<img class="tfp_lrg_img" src="(.*)" alt=', page)
    if match:
        img_url = NEWSEUM_IMG + os.path.abspath(match.group(1))
        print "Saving the image for %s" % (city)
        image = urllib2.urlopen(img_url).read()
        open(os.path.basename(match.group(1)), "wb").write(image)

def main():
    """
    Main code goes here
    """
    for city in CITIES:
        fetchNewseumImage(city)

if __name__ == '__main__':
    main()

The source is also hosted here at github

Hygrosens manufactures a number of sensors for measuring temperature, humidity, light level, pressure. Their devices use a common serial data format for a wide variety of sensors, include 1-wire sensors from Dallas. This library reads the output from Hygrosens devices and passes it to a calling function as a hash. I have included an example that outputs the readings in human readable format, and another that stores the readings into a MySQL database.