PyQGen

• Author: Giuseppe Lipari
• Email : giuseppe.lipari@univ-lille.fr

PyQGen is a command line script to generate a randomized list of questions taken from a orgmode file. It can be used to prepare exams for large classes.

What you need in addition to PyQGen :

• Emacs with org-mode;
• A LaTeX environment.

Overview

The original list of questions must be redacted according to the org-mode format. The first level heading in this file represent groups of questions; the second level headings represent the questions ; deeper level headings represent solutions.

An example of database of questions is available here.

PyQGen produces an org-mode file which contains the exams. This can later be transformed into a PDF file via LaTeX.

PyGen also produces an excel sheet for correction and grading.

Installing

I recommend installing PyQGen using a virtual environnement and pip. For example :

virtualenv pyqgenenv
source ./pyqgenenv/bin/activate
pip install pyqgen

PyQGen depends on

• orgparse to parse the org-mode file,
• openpyxl to produce the excel file for grading.

Pip automatically takes care of the dependencies.

Command line options

The command is :

pyqgen output [OPTIONS] 

where output the generated file that contains the exams. The following options are possible :

• -h, --help shows the help message
• -d DB, --db DB specifies an input file. By default, this is equal to file db.org.
• -t TITLE, --title TITLE Specifies the title of each exam sheet (default: "Examen")
• -i IFILE, --ifile IFILE Text file containing the instructions to be printed on each exam (default: none)
• -n NCOPIES, --ncopies NCOPIES Number of exams to generate (default: 1)
• -g [NG ...], --ng [NG ...] Number of questions per group (default: [1, 1, 1]). Therefore, the default assumes that the input file defines 3 question groupes, and for each exam it will randomly select one question per group. Make sure that you specify at least one number per each group (see example below).
• -e HEADER, --header HEADER Org-mode file header. This is used to personnalize the output style. Typically, you can specify the size of the sheet (using the latex package geometry), the font, the font size, etc.

Examples

Go in directory examples/basic. The database of questions is in file db.org. Please open the file to familiarize with the way it is structured. In this example, there are 3 question groups (first heading) each one containing questions, exercises and solutions.

Solutions are third level heading tagged with the :solution: tag.

For each question, we can optionally specify three properties:

• CATEGORY is a comma-separated list of categories, each one represents a set of questions. The program never generates questionnaires with two questions belonging the same category. This can be useful to avoid having two similar questions in the same exam.

• NUM_RESP and NUM_CORRECT are used for feedback: the first one is the number of time the question has been answered in a questionnaire; the second one is the number of correct answers to this question. The rate NUM_CORRECT/NUM_RESP represents an indication of how easy is the question. THIS IS AN EXPERIMENTAL FEATURE, not yet fully implemented.

To compile the database file into a pdf, run

./compile.sh db.org

To generate the questionnaire, you can run one of the following commands :

• pyqgen out.org -n 30

  This generates two files: out.org contains 30 copies of the questionnaire, and out.xls contains the grading sheet.
  To compile the questionnaire into a PDF file, run compile.sh out.org. Every instance of the questionnaire contains 3 questions, one per question group.

• pyqgen out.org -n 30 --ng 2 1 1

  This command generates questionnaires consisting of 4 questions each: 2 taken from the first group, 1 from the second group, and 1 from the third group.

After generating the questionnaire, you can run

./compile.sh

to get the out.pdf file in output.