Open gwaybio opened 6 years ago
Given the short time frame, I am leaning more towards option 1 for 2018. We will at least have a working product this way
If the .local approach is a 10 minute fix for them, then we'd love to do it that way. If it's not going to be ready (realistically) today, then I think we should do option 1. With the work that you've already done, we should be able to ask them for the latest versions of the relevant packages that you've dug up.
Either way, communicating the strong desire for the .local
option is probably helpful for next year :)
@gwaygenomics What did you decide on this? Wondering how we should triage the issues you have for the 2018 Milestone.
yeah... add to 2019 milestone. I requested the .local
feature to the cocalc admin and I think they are working on it.
I reverted the first couple modules back to python (sagemath)
kernel, but there are more I need to revert back (from python (GCB535)
). @kelsj and I need to make sure we don't push out notebooks with this kernel - they cause errors that could be pretty confusing
Ok - can you update the milestones for whichever issues are affected?
I sent a new ticket to catch up on this. It's #6071 in zendesk for them. I'll update when I hear more.
We corresponded with you last year about this. Some of your responses are tracked in this github issue: https://github.com/greenelab/GCB535/issues/161
I'm curious whether or not you've had a chance to implement a solution that would help us push an environment out to the students.
Thanks! Casey
A feature that could help has been added. We could modify the first assignment to include the code necessary to setup the environment: https://doc.cocalc.com/teaching-tips_and_tricks.html#run-terminal-command-in-all-student-projects
Here is the full reply:
Hi, Casey.
We don't have specific support for e.g. distributing custom kernels in ~/.local. But we have added a feature that may help. If you use Handouts to distribute files, you can then run a command to move/copy/etc. in each student project. Please see "Run Terminal command in all student projects" in the User Manual: https://doc.cocalc.com/teaching-tips_and_tricks.html#run-terminal-command-in-all-student-projects
Hope this helps.
Hal
It appears that setting up a custom conda environment for students is more complicated than I initially thought. See #155 for details leading up to this issue.
The folks at cocalc have been responsive to my support ticket and I have gotten permission from them to post their responses below.
Summary
To summarize, it appears that we have a couple of options:
Python 2 (SageMath)
orR (SageMath)
) and ask for packages when neededsetup_environment.sh
script to be run as one of the first modules in the course.tagging @cgreene and @bvoight for discussions/decisions - thanks!
CoCalc Ticket:
My Initial Question
First Response
Second Response