ilebedev / py_web_gui

Tiny RPC framework that enables Python to use a web view in a browser session as a GUI.

py_web_gui

Tiny RPC framework that enables Python to use a web view in a browser session as a GUI.

Prerequisites

In order to use the s04 infrastructure, you will need a computer with a linux-like terminal (I haven't tried building labs on Windows, but did make a reasonable effort at writing portable infrastructure). You will need git and zip. You will also need python 2.7.

While you won't need much to build the business logic of an assignment, you will require a PDF tool to build the assignment documents. In this repository, I do not include any complete assignment, but if you choose to use this code in its entirety, the following are prerequisites for building PDFs:

Prerequisites for building PDF assignment documents

You will need gem (provided by a recent version of ruby), and wkhtmltopdf installed and available at a command line.

• Download and install wkhtmltopdf here. Don't use the one hosted by Ubuntu, it is missing some important components.
• Install the required gems: gem install redcarpet github-markup.

Building labs

To build a completed lab (create the .zip archives for funprog (as provided by the igor project), and a solution), use the make_assignment.py script. For example:

./make_assignment.py labs/gas

will create .lab_1.zip, .analyzer_1.zip, and solution.lab_1/ - the complete set of files to be ever shipped to students.

Running the lab

A solution can be tested via ./test.py, or python test.py.

Run the UI via ./server.py, or python server.py, and use your favorite web browser (we tested everything on Chrome) to visit localhost:8000. If there are any problems, you can change the port number by editing server.py.

Creating new labs

Each assignment must contain:

• config.json, which tells the framework everything it needs to know in order to biuld the assignment.
• a sub-folder cases, containing the .in and .out files with contiguous numeric names beginning with 1.
• a sub-folder resources, containing anything that is needed by both the autograder and the UI (data sets, for example).
• verifier.py, which implements verify( result, input_data, gold ): comparing result produced by the student code against gold, which is the known good solution (the .out file). input_data (the .in file) is also available. If solutions are not unique, verifier should check an invariant which determines whether or not the solution is correct.
• wrapper.py, which implements run_test( input_data ): a function that knows how to invoke student code given input_data (the .in file).
• empty.py: the file students will work in to produce a solution.
• solution.py: a staff solution for the assignment.

Any documents with the assignment (the assignment PDF, for example) can be written in markdown, which is also used to format doucments on GitHub. Each markdown document to be built should be added to the markdown key in config.json.

Furthermore, if the assignment has a web-hosted UI (enabled via a flag in config.json), the UI belongs in a sub-folder ui.