After more than 1/4 century, my email address is going to die. ☹
I'm "feeling homeless" now. ☹ Suggestions for privacy-oriented email hosts
are highly welcome.
This is because there are clashes in course and specialization names.
For example, "machine-learning" is both a course and a specialization
name. These are different courses from different universities (!).
Bugfixes:
- Switch to newer API for syllabus and lecture retrieval (#665, #673, #634)
Features:
- You can now download specializations: the child courses will be
downloaded automatically
I wish I had time to track down the real URL that's preventing TLS
negotiation, but this at least gets coursera-dl working with its own
downloader again.
`clean_filename` wasn't used; it's now used before making directories
or write files under the notebook folder.
`clean_filename` is improved a little, to handle more Windows-specific
edge cases. And the existing tests were updated accordingly.
The structure is very similar to all other text instructions but peer
assignment instructions should be retrieved using the following API method:
onDemandPeerAssignmentInstructions.v1
fix#650
Coursera Notebooks is a new section in some courses that includes
some interactive files that are note in the resources. According
to the Python Jupyter format, it is mandatory to get them on the
go because the session id (it changes on each request).
Features:
- Default arguments are loaded from `coursera-dl.conf` file
- Added option `--mathjax-cdn <MATHJAX_CDN>` to specify alternative MathJax CDN
- Added support for Resources section
Features:
- Add `--download-delay` option that adds a specified delay in seconds
before downloading next course. This is useful when downloading many
courses at once. Default value is 60 seconds.
- Add `--only-syllabus` option which is when activated, allows to skip
download of the course content. Only syllabus is parsed.
- Add support for `reflect` and `mcqReflect` question types in quizzes.
- Courses that encountered an error while parsing syllabus will be listed
in the end of the program execution, after all courses have been
processed (hopefully, downloaded). This helps skip vast output and easily
see which courses need user's attention, e.g. enrollment, session
switching or just patience until the course start date.
Bugfixes:
- Locked programming assignments in syllabus used to crash coursera-dl.
Now the script goes on parsing syllabus and skips locked assignments.
- Add missing import statement to playlist generation module
This change makes the script show courses where an error
during syllabus parsing occurred. Printing such courses
in the end is helpful to bring attention when downloading
multiple courses.
related #573
reflect generates the same narrow edit field as for numerical answers.
Yet it does not look as on Coursera site, I think it's enough
to indicate the intention. mcqReflect is basically the same as just mcq.
related #573
This option makes the script sleep for N seconds before
downloading next course. Default value is 60 seconds.
This option proved useful when downloading many courses at
once. Otherwise Coursera API starts to respond with errors
(cannot login).
related #573
