Coursera Downloader

Coursera is arguably the leader in massive open online courses (MOOC) with a selection of more than 300 classes from 62 different institutions as of February 2013. Generous contributions by educators and institutions are making excellent education available to many who could not afford it otherwise. There are even non-profits with "feet on the ground" in remote areas of the world who are helping spread the wealth (see the feedback below from Tunapanda).

This script makes it easier to batch download lecture resources (e.g., videos, ppt, etc) for Coursera classes. Given one or more class names and account credentials, it obtains week and class names from the lectures page, and then downloads the related materials into appropriately named files and directories.

Why is this helpful? A utility like wget can work, but has the following limitations:

Video names have a number in them, but this does not correspond to the actual order. Manually renaming them is a pain.
Using names from the syllabus page provides more informative names.
Using a wget in a for loop picks up extra videos which are not posted/linked, and these are sometimes duplicates.

DownloadThemAll is another possibility, but this script provides more features such as appropriately named files.

This work was originally inspired in part by youtube-dl by which I've downloaded many other good videos such as those from Khan Academy.

Features

Intentionally detailed names, so that it will display and sort properly on most interfaces (e.g., MX Video on Android phone).
Regex-based section (week) and lecture name filters to download only certain resources.
File format extension filter to grab resource types you want.
Login credentials accepted on command-line or from .netrc file
Core functionality tested on Linux, Mac and Windows.

Instructions

coursera-dl requires Python 2 (2.6 or newer) or Python 3 (3.2 or newer) and a free Coursera account enrolled in the class of interest. Note: You must already have (manually) agreed to the Honor of Code of the particular courses that you want to use with coursera-dl.

Install any missing dependencies.

We strongly recommend that you consider installing Python packages with pip, as in it is the current preferred method. If you are using pip, you can directly install all the dependencies from the requirements file using pip install -r requirements.txt.

Installing dependencies on your own

Warning: This method is not recommended unless you know what you are doing. Before filing bug reports, please check that the versions of your modules are those recommended in the requirements.txt file.

You may choose to install the dependencies yourself, but our users had issues that not all resources (videos etc.) were downloaded with versions of the dependencies different than those listed in the requirements.txt file.

In any case, you may want to install:

Beautiful Soup 4: Required. See also html5lib below.
- Ubuntu/Debian: sudo apt-get install python-bs4
- Mac OSX + MacPorts: sudo port install py-beautifulsoup4
- Other: pip beautifulsoup4
Argparse: Required (but you only need to install with Python 2.6)
- Ubuntu/Debian: sudo apt-get install python-argparse
- Other: pip argparse
requests: Required.
- Ubuntu/Debian: sudo apt-get install python-requests
- Mac OSX + MacPorts: sudo port install py-requests
- Other: pip requests
six: Required.
- Ubuntu/Debian: sudo apt-get install python-six
- Mac OSX + MacPorts: sudo port install py27-six
- Other: pip six
html5lib: Not required, but recommended for parsing pages.
- Ubuntu/Debian: sudo apt-get install python-html5lib
- Mac OSX + MacPorts: sudo port install py-html5lib
- Other: pip html5lib
easy_install: Only necessary if not using prepackaged dependencies. Also, pip supersedes it.
- Ubuntu/Debian: sudo apt-get install python-setuptools

Again, make sure that you have the versions mentioned in the file requirements.txt (later versions may be OK).

On Mac OSX using MacPorts, the following may be used:

port
> select --set python python27
> install py-beautifulsoup
> install py-argparse
> install py-setuptools
> install py-requests
> install py27-six

Create an account with Coursera

If you don't already have one, create a Coursera account and enroll in a class. See https://www.coursera.org/courses for the list of classes.

Running the script

Run the script to download the materials by providing your Coursera account (e.g., email address), password (or a ~/.netrc file), the class names

General:                     coursera-dl -u <user> -p <pass> modelthinking-004
Multiple classes:            coursera-dl -u <user> -p <pass> saas historyofrock1-001 algo-2012-002
Filter by section name:      coursera-dl -u <user> -p <pass> -sf "Chapter_Four" crypto-004
Filter by lecture name:      coursera-dl -u <user> -p <pass> -lf "3.1_" ml-2012-002
Download only ppt files:     coursera-dl -u <user> -p <pass> -f "ppt" qcomp-2012-001
Use a ~/.netrc file:         coursera-dl -n -- matrix-001
Get the preview classes:     coursera-dl -n -b ni-001
Specify download path:       coursera-dl -n --path=C:\Coursera\Classes\ comnetworks-002

Maintain a list of classes in a dir:
  Initialize:              mkdir -p CURRENT/{class1,class2,..classN}
  Update:                  coursera-dl -n --path CURRENT `ls CURRENT`

On *nix platforms, the use of a ~/.netrc file is a good alternative to specifying both your username and password every time on the command line. To use it, simply add a line like the one below to a file named .netrc in your home directory (or the equivalent, if you are using Windows) with contents like:

machine coursera-dl login <user> password <pass>

Create the file if it doesn't exist yet. From then on, you can switch from using -u and -p to simply call coursera-dl with the option -n instead. This is especially convenient, as typing usernames and passwords directly on the command line can get tiresome (even more if you happened to choose a "strong" password).

Troubleshooting

It is important that:

When reporting bugs against coursera-dl, please don't forget to include enough information so that you can help us help you:
- Is the problem happening with the latest version of the script?
- What operating system are you using?
- Do you have all the recommended versions of the modules? See them in the file requirements.txt.
- What is the course that you are trying to access?
- What is the precise command line that you are using (feel free to hide your username and password with asterisks, but leave all other information untouched).
- What are the precise messages that you get? Please, copy and paste them. Don't reword the messages.
Make sure the class name you are using corresponds to the resource name used in the URL for that class: https://class.coursera.org/<CLASS_NAME>/class/index
Note that many courses (most, perhaps?) may remove the materials after a little while after the course is completed, while other courses may retain the materials up to a next session/offering of the same course (to avoid problems with academic dishonesty, apparently).

In short, it is not guaranteed that you will be able to download after the course is finished and this is, unfortunately, nothing that we can help you with.
Make sure you have installed and/or updated all of your dependencies according to the requirements.txt file as described above.
One can export a Netscape-style cookies file with a browser extension (1, 2) and use it with the -c option. This comes in handy when the authentication via password is not working (the authentication process changes now and then).
If results show 0 sections, you most likely have provided invalid credentials (username and/or password in the command line or in your .netrc file).
For courses that have not started yet, but have had a previous iteration sometimes a preview is available, containing all the classes from the last course. These files can be downloaded by passing the -b parameter.

If you are using Beautiful Soup 4, make sure you have installed html5lib:

  $ python
  >>> import html5lib
  >>> print(html5lib.__version__)
  1.0b2

If you get an error like Could not find class: <CLASS_NAME>:
- Verify that the name of the course is correct. Current class names in coursera are composed by a short course name e.g. class and the current version of the course (a number). For example, for a class named class, you would have to use class-001, class-002 etc.
- Second, verify that you are enrolled in the course. You won't be able to access the course materials if you are not officially enrolled and agreed to the honor course via the website.
If:
- You get an error when using -n to specify that you want to use a .netrc file and,
- You want the script to use your default netrc file and,
- You get a message saying coursera-dl: error: too few arguments
Then you should specify -- as an argument after -n, that is, -n -- or change the order in which you pass the arguments to the script, so that the argument after -n begins with an hyphen (-). Otherwise, Python's argparse module will think that what you are passing is the name of the netrc file that you want to use. See issue #162.

Feedback

I enjoy getting feedback. Here are a few of the comments I've received:

"Thanks for the good job! Knowledge will flood the World a little more thanks to your script!"
Guillaume V. 11/8/2012
"Just wanted to send you props for your Python script to download Coursera courses. I've been using it in Kenya for my non-profit to get online courses to places where internet is really expensive and unreliable. Mostly kids here can't afford high school, and downloading one of these classes by the usual means would cost more than the average family earns in one week. Thanks!"
Jay L., Tunapanda 3/20/2013
"I am a big fan of Coursera and attend lots of different courses. Time constraints don't allow me to attend all the courses I want at the same time. I came across your script, and I am very happily using it! Great stuff and thanks for making this available on Github - well done!"
William G. 2/18/2013
"This script is awesome! I was painstakingly downloading each and every video and ppt by hand -- looked into wget but ran into wildcard issues with HTML, and then.. I came across your script. Can't tell you how many hours you've just saved me :) If you're ever in Paris / Stockholm, it is absolutely mandatory that I buy you a beer :)"
Razvan T. 11/26/2012
"Thanks a lot! :)"
Viktor V. 24/04/2013

Contact

Post bugs and issues on github. Send other comments to John Lehmann: first last at geemail dotcom or @jplehmann

ans-/coursera