This library aims to perform two functions:
- Parse the XML code for an edX course, loading it into python objects
- Validate the objects for errors
Based on this, two scripts are provided that leverage the library:
edx-cleaner
constructs an error report, course tree and course statisticsedx-reporter
constructs a LaTeX file representation of the course structure
Version 0.1.3
Copyright (C) 2018-2019 Jolyon Bloomfield
This package may be installed from PYPI using pip install olxcleaner
. It requires python 3.6 or later.
Clone this repository, and set up a virtual environment for python 3.6 or later. Run pip install -r requirements.txt
to install the libraries, followed by pytest
to ensure that all tests are passing as expected.
Used to validate OLX (edX XML) code. This is a very light wrapper around the olxcleaner library, but exposes all of the functionality thereof.
Basic usage: run edx-cleaner
in the directory of the course you want to validate.
Command-line options:
edx-cleaner [-h]
[-c COURSE]
[-p {1,2,3,4,5,6,7,8}]
[-t TREE] [-l {0,1,2,3,4}]
[-q] [-e] [-s] [-S]
[-f {0,1,2,3,4}]
[-i IGNORE [IGNORE ...]]
-h
: Display help.-c
: Specify the course file to analyze. If not specified, looks forcourse.xml
in the current directory. If given a directory, looks forcourse.xml
in that directory.-p
: Specify the validation level you wish analyze the course at:- 1: Load the course
- 2: Load the policy and grading policy
- 3: Validate url_names
- 4: Merge policy data with course, ensuring that all references are valid
- 5: Validate the grading policy
- 6: Have every object validate itself
- 7: Parse the course for global errors
- 8: Parse the course for detailed global errors (default)
-t TREE
: Specify a file to output the tree structure to.-l
: Specify the depth level to output the tree structure to. Only used if the-t
option is set. 0 = Course, 1 = Chapter, 2 = Sequential, 3 = Vertical, 4 = Content.-q
: Quiet mode. Does not output anything to the screen.-e
: Suppress error listing. Implied by-q
.-s
: Suppress summary of errors. Implied by-q
.-S
: Display course statistics (off by default). Overridden by-q
.-f
: Select the error level at which to exit with an error code. 0 = DEBUG, 1 = INFO, 2 = WARNING, 3 = ERROR (default), 4 = NEVER. Exit code is set to1
if an error at the specified level or higher is present.-i
: Specify a space-separated list of error names to ignore. See Error Listing.
The olxcleaner library includes modules that parse a course into python objects. This can be useful if you want to scan a course to generate a report. We exploit this in edx-reporter
to generate a LaTeX report of course structure.
Basic usage: run edx-reporter
in the directory of the course you want to generate a report about.
Command-line options:
edx-reporter.py [-h]
[-c COURSE]
[-u]
[> latexfile.tex]
-h
: Display help.-c
: Specify the course file to analyze. If not specified, looks forcourse.xml
in the current directory. If given a directory, looks forcourse.xml
in that directory.-u
: Include url_names for verticals.> latexfile.tex
: Output the report to a file.
If you get an error like Character cannot be encoded into LaTeX: U+FEFF - `'
, then you have some bad unicode in your display_name
entries. Look through the LaTeX output for {\bfseries ?}
, which is what that character is converted into.
Once you have generated a latex file, you can compile it into a PDF file by running pdflatex latexfile.tex
. Note that the latex file can be modified with any text editor; its format should be self-explanatory.
The workhorse of the library is olxcleaner.validate
, which validates a course in a number of steps.
olxcleaner.validate(filename, steps=8, ignore=None)
filename
: Pass in either the course directory or the path ofcourse.xml
for the course you wish to validate.steps
: Choose how many validation steps you wish to perform:- 1: Load the course
- 2: Load the policy and grading policy
- 3: Validate
url_name
s - 4: Merge policy data with course, ensuring that all references are valid
- 5: Validate the grading policy
- 6: Have every object validate itself
- 7: Parse the course for global errors
- 8: Parse the course for global errors that may be time-consuming to detect
ignore
: A list of error names to ignore
Returns EdxCourse
, ErrorStore
, url_names
(dictionary {'url_name': EdxObject}
, or None
if steps < 3
)
See examples of how to use olxcleaner.validate
and the objects it returns in olxcleaner.entries
.