Use new skan.summarize(separator="_") #986
Assignees
Labels
Milestone
Comments
|
Discovered tests were failing on #985 with this new Its related to Numba and type errors. Will see if I can fix upstream before making these changes locally. |
|
This should await the release of |
|
skan-0.12.1 has been released but does not include the proposed fix as its awaiting me extracting the data that caused the error and writing a test ensure there is no regression in the future. Might try doing that next week as we can then unpin the version we depend on and remove the substitution that is required to be done manually. |
|
skan-0.12.2 published which includes the bug fix I made to address the above issue. This means we should be able to...
|
ns-rse
added a commit
that referenced
this issue
Dec 16, 2024
Closes #986 With the release of `skan-0.12.2` we can remove the code which converts `-` in the variable names of Pandas DataFrames to instead use `_` which allows the use of "dot-notation" when referring to the variables of data frames. We therefore unpin the `skan` version in `pyproject.toml`, explicitly state the `separator="_"` argument to `skan.summarize()` and adapt the hard-coded variable names appropriately. As noted in the [release notes](https://github.com/jni/skan/releases/tag/v0.12.2) the default will be `_` from `v0.13.0` and the flag allows backwards compatibility but since we have gone with the use of `_` in variable names it will do no harm leaving these in place (although they could be removed when `skan-v0.13.0` is released).
ns-rse
added a commit
that referenced
this issue
Dec 16, 2024
In addressing #986 I tested in a fresh virtual environment. This pulled in [matplotlib-3.10.0](https://github.com/matplotlib/matplotlib/releases/tag/v3.10.0) which was released 2024-12-14 (two days ago!). Our dependency on [topoly](topoly.cent.uw.edu.pl/documentation.html) doesn't work with this new version because of the removal of `inset_location.InsetPosition` in favour of `inset_axes`. This has been reported to the Topoly developers, see [ImportError: cannot import name 'InsetPosition' from 'mpl_toolkits.axes_grid1.inset_locator' · Issue #5 · ilbsm/topoly_tutorial](ilbsm/topoly_tutorial#5) and hopefully won't take long to resolve. For now though we pin our dependency `matplotlib~=3.9.4` and all tests pass.
2 tasks
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
We added a line in disrordered_tracing() to substitute the
-in column names of dataframes produced by Skan.A new version of skan has been released (19hrs ago) which will include the minor contribution I made of converting all variable names from
-to_(see release notes and linked pull request).We should, until
_becomes the default, setskan.summarize(..., separator="_").The text was updated successfully, but these errors were encountered: