UnsatisfiableError with a viz package

RenkuLab
1

Hi there! Just running into a Python dependency issue trying to use Vega from https://github.com/vega/ipyvega Anyone dealt with this kind of issue before? How does one pin versions in the requirements file, anyway?

UnsatisfiableError: The following specifications were found to be incompatible with a past
explicit spec that is not an explicit spec in this operation (notebook):

  - vega -> notebook[version='>=4.2']
2

Hi @loleg, thanks for raising this problem - it looks like the latest conda is installed and potentially causing problems. Can you try installing with pip?

3

It looks like the conda version is pinned somewhere - I’m trying to locate where that is. Could be in the base image from the jupyter docker stack that we are using…

1 Like
4

Sure: I get a slightly different looking but essentially same error at https://renkulab.io/gitlab/oleg/swiss-hospitals/-/jobs/25492

5

I think I’ve gotten to the root of it - the base jupyter images we’re using lock down the conda channels and enforce strict channel order for package dependency resolution. Apparently this has caused conda to completely melt down in this case. I removed the line that enforces the channel order in this Dockerfile and it seems to work.

1 Like
6

This did fix the issue :tada: thanks @rrrrrok

My next question: is there any way I can create a merge request in GitLab from your repository to ours? Or let me put it another way - where do you have guidelines for people to collaborate, sync their forks, etc.?

7

Yes you can initiate a merge request from the source repo to your fork. Alternatively, you can start a session and do the following in the terminal:

git remote add upstream https://renkulab.io/gitlab/covid-19/covid-19-public-data.git
git fetch upstream 
git merge upstream/master

Sadly we lack the documentation for this at the moment - I need to add it at least to the readme of the covid-19 public data project - is that the fork you want to update? In general, if you want to frequently pull the data in from the source repo it’s best if you don’t work on the same notebooks and data directories in your fork so you can avoid merge conflicts.