You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
There are currently hard pins of bokeh and jinja2, which make it very difficult to install tedana alongside other tools that may require more recent versions. For example, installing jupyterlab into an environment with tedana installed produces:
ERROR: pip's dependency resolver does not currently take into account all the packages that are installed. This behaviour is the source of the following dependency conflicts.
tedana 23.0.1 requires jinja2==3.0.1, but you have jinja2 3.1.2 which is incompatible.
This is not currently a problem for me, as I am not using these features of tedana, but using ==, or < pins tends to make it difficult to resolve dependencies.
It looks like these were added in #793 and #870, which are 2 and 1.5 years old, respectively. It is not uncommon for tools to only support dependencies released within the last 2 years or so, so it is going to become increasingly difficult to install tedana into the same environment as other tools in the near future.
A little related to #934, but not exactly the same.
Additional Detail
Next Steps
IIRC, there was a weirdness with Jinja 3.0.2, so possibly that pin can just be removed at this point (or updated to jinja2 != 3.0.2.
I don't know anything about bokeh and how hard it will be to accommodate a more recent version.
The text was updated successfully, but these errors were encountered:
I just updated to jinja2=3.1.2 and I'm getting the same error so we cannot just skip a version or two.
An issue that provides more detail is #869. The jinja2 & bokeh issues are linked. bokeh was locked on older versions of jinja2 & it looks like bokeh removed that limit about a year ago so we should able to remove our jinja2 limit if we update bokeh.
The person who worked on the bokeh parts of the tedana reports moved on to other things so we'd benefit from someone with more bokeh experience to figure out what's still holding back a bokeh upgrade & fix that.
I have worked a bit on the bokeh side of things in our reports in the past. I can try to have a look and see if there is anything we can do with these dependencies.
Summary
There are currently hard pins of bokeh and jinja2, which make it very difficult to install tedana alongside other tools that may require more recent versions. For example, installing jupyterlab into an environment with tedana installed produces:
This is not currently a problem for me, as I am not using these features of tedana, but using
==
, or<
pins tends to make it difficult to resolve dependencies.It looks like these were added in #793 and #870, which are 2 and 1.5 years old, respectively. It is not uncommon for tools to only support dependencies released within the last 2 years or so, so it is going to become increasingly difficult to install tedana into the same environment as other tools in the near future.
A little related to #934, but not exactly the same.
Additional Detail
Next Steps
IIRC, there was a weirdness with Jinja 3.0.2, so possibly that pin can just be removed at this point (or updated to
jinja2 != 3.0.2
.I don't know anything about bokeh and how hard it will be to accommodate a more recent version.
The text was updated successfully, but these errors were encountered: