bgw / century Goto Github PK

We currently almost any testing mechanisms. This is primarily because of the difficulty of testing such an organic system, where results often change based on who is using the library (like when reading a schedule). Still, testing is important. We need a plugin for the the lib.browser.Browser class, overriding the load_page method to load in static, local pages to make testing a little easier.

Registrar and ISIS-based task modules will likely have some similar functionality, and as much stuff as possible should be moved into a courses module. Some work has been done here, but more work should be done towards this, to make it as good as possible.

The HTTP/1.1 protocol supports sending gzipped content to speed up page loads. We should take advantage of this. Ideas on where to start can be found via various sources online. This should be done as a urllib handler if at all possible, allowing maximum portability.

Make a plugin for lib.browser.Browser to wait either a fixed amount of time, or a variable amount of time (determined by a function call), before loading each page, as to be nicer on UF's servers if we are loading a large number of pages.

It'd be good to have the ability to scrape the page you see after searching for a course number in the "Registration" section, to automate checking for openings in classes.

There's some documentation included in the source for the lib.tasks.phonebook module and its submodules, but it should be cleaned up and included in the Sphinx documentation.

Add support to lib.browser.plugins.Pluggable to dynamically unload plugins.

There's a very weak base of a module in lib.tasks.registrar. This should be fleshed-out and made functional.

PyPi is the standard package index for Python packages. Century should be packaged with setuptools and made available for easy installation through PyPi. Having Century available in other package types, like deb and rpm would be nice, but PyPi should be the priority.

Right now, lib.browser.plugins.keepalive only supports HTTP pages. On HTTPS, Keep-Alive support is disabled. This Keep-Alive support should be added for HTTPS pages, especially since HTTPS is probably the most common protocol we use, and could use the speedup the most.

We need some sort of introductory manual for Century, getting new developers started with simple tutorials and small projects, requiring minimal amounts of code to interest people. The target audience should be someone with a low-to-moderate knowledge of Python. Some features like decorators, if discussed, should be explained briefly, or a link explaining how they work should be included.

For many classes, the Sphinx documentation contains both properties, and the getters (and in some cases, setters) for them. The Sphinx documentation files should be modified to exclude (simple) getter and setter methods. They'll still be there, but in most cases, it just adds clutter and confusion to the documentation.