Check model coords for unknown shapes when building predictive models #413
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
AlexAndorra
approved these changes
Jan 12, 2025
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Outdated suggestions cannot be applied.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.
Closes #412
When a user defines a time-varying parameter, he has to give it shape None when building the symbolic graph. This is because the time dimension, unlike all other dimensions, has to remain flexible.
A consequence of this is that the
.type.shapeinformation will have aNone. This was assumed to always be known by thePyMCStateSpacemodel class, and was used when makingpm.FlatRVs in predictive models (sample_conditional,sample_unconditional,impulse_response_function, andforecast). PyMC evidently doesn't let you passNoneas a shape to an RV.Since
PyMCStateSpacestores the coords used to fit the model, we can just check those as a fallback to get the expected shapes of RVs. This PR does that. Tasks that require changing the shape of the time dimension will still fail (e.g.forecast), but that is always going to be true with exogenous time-varying parameters.