WIP 0.8.0 #106
Merged
WIP 0.8.0 #106
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Implement loglik ratio test analogous to the one implemented in lme4
Make output closer to lme4
Make sure npar are correct and other values are in the same order
AICTab in summary of an lmer fitted by REML gives the deviance, not the AIC.
For the moment, we use a static dump of the R output, as rpy2 seems to clash with pytest
|
I'd love some help with anyone willing to take a look at making the CI work (i.e. wrestling with Once the test suite passes and/or bugs are fixed I'll cut a new release |
…d remove global conversion activators
…verting logits to probs in Lmer. Fixes #78"
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Overview
This PR originally started to incorporate changes made by @dramanica, but got caught up in
rpy2andconda-buildhell. That gave me enough time to clean up and fix a few other issues.I've also overhauled the CI setup such that all PRs and pushes to main/master run a "lighter" CI suite which just performs a minimal install of
pymer4in a new conda environment. This is feasible usingconda create -n pymer4_env python=3.8 'r-lmerTest' r-emmeans' 'rpy2'and then proceeding with a normalpipinstall (with requirements). This has the benefit of decoupling the test and build suites such that it's possible for ensure that the library is functioning properly regardless ofcondadifficulties.condawas supposed to make things easier, but also now I find myself using tools like act to reproduce the full CI setup locally just to test building, while building locally works just fine...rabbit holes all the way down.That being said I'm planning on skipping 0.7.9 and only releasing 0.8.0 given the large compatibility changes and new model type (see below)
Summary of Changes
Fixes
Description
pandasandrpy2depslocalconvertereverywhere to translate between R and Python objects facilitated with the newbridgemodule. This was made possible by pulling the solution from this commit now that the upstream issue has been fixedLmmodels now supportfamily='binomial'for logistic regression. Currently on the most basic model is implement with no permutation testing, bootstrapping, or predictionBlocking Build Issue
Currently several of the GA runners that try to build pymer4 are failing for this reason. There seems to be a pending upstream change but it appears to be stuck.
Forcing a higher numpy version didn't work either as the conda builder just isn't picking it up on GA (but it does locally).
I think the next step is to try to use conda mambabuild...