webclient.txt
+--

Assignment: choose one or both or all three. I suspect 2 is a lot harder, but 
then you would have learned a lot about the GitHub API and JSON.

1. File downloader using screen scraping: write a script that scrapes
our Python Winter 2012 course page for the URLs of all the Python
source files, then downloads them all.

2.  File downloader using web API: same as 1, but find the Python source
files using the GitHub API.  Hints: you need to look in the repo at
https://github.com/jon-jacky/uw_python/tree/gh-pages, not the web
pages.  Git calls the files in a repository "blobs", see the "get all blobs"
example about halfway down this page: http://develop.github.com/p/object.html

3. Write any web client of your choosing.  The point here is to create
something that you find interesting or useful -- don't worry about
using a lot of tools and technique.  When you have something to show,
post the code on github and email me about it.


+--

Warning!


When testing any automated web client, take care not to overdo it --
don't overload the server.  In 1. and 2. above, the server is github;
each run of your program imposes a similar load as cloning the
repository.  Test your program progressively: first, confirm you are
scraping the Python URLs correctly before testing any code to download
them.  Then, test downloading from a local cloned (by hand) copy of
the repository, using file: urls (not http:).  When that's working,
try downloading from github.  See the github terms of service at
http://help.github.com/terms-of-service/