-
Notifications
You must be signed in to change notification settings - Fork 45
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Moseq Pipeline #1056
base: master
Are you sure you want to change the base?
Moseq Pipeline #1056
Conversation
…t for Pose-dependent analysis
Check out this pull request on See visual diffs & provide feedback on Jupyter Notebooks. Powered by ReviewNB |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Took a first pass on the behavior subpackage - Can continue review tomorrow
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Looking good so far!
comments=pos_obj.comments, | ||
description=pos_obj.description, | ||
) | ||
if isinstance(pos_nwb["dlc_position"], pd.DataFrame): |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I think I would migrate these to a separate method to clean up the indenting, something like declare_pos_analysis_file
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
@CBroz1, reorganized this make function. Let me know if it looks cleaner to you
Co-authored-by: Chris Broz <[email protected]>
I've now merged in the main branch and have tested it working with a dependency on |
A user will need a DLC project in place to use MoSeq, right? I think I might rename our 'dlc' optional dependency to 'position' and moseq dependencies in there - not sure which I prefer across the following:
|
Co-authored-by: Chris Broz <[email protected]>
I may lean towards keeping them separate. You don't strictly need DLC in your environment to run the moseq pipeline.
|
View / edit / reply to this conversation on ReviewNB edeno commented on 2025-01-20T20:30:33Z Line #1. from spyglass.behavior.core import PoseGroup Should be:
|
View / edit / reply to this conversation on ReviewNB edeno commented on 2025-01-20T20:30:33Z Line #1. from spyglass.behavior.moseq import ( should be:
|
View / edit / reply to this conversation on ReviewNB edeno commented on 2025-01-20T20:30:34Z Line #5. MoseqModel().populate(model_key) This didn't work for me on testing but it also didn't give an informative error message as to why. Any thoughts? samuelbray32 commented on 2025-01-27T23:02:28Z When I deleted and ran this again I had an issue related due to the project path. I changed the code to make it unique prior to creation of the moseq model
|
View / edit / reply to this conversation on ReviewNB edeno commented on 2025-01-20T20:30:35Z Line #2. table.analyze_pca() This failed for me as well:
|
View / edit / reply to this conversation on ReviewNB edeno commented on 2025-01-20T20:30:36Z Line #3. from spyglass.behavior.moseq import MoseqSyllableSelection, MoseqSyllable
|
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I ran through the notebook and ran into a couple of problems. I was unable to run the moseq algorithm I think but it was unclear to me why.
"jax-moseq", | ||
"keypoint-moseq", | ||
] | ||
|
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Update the omit
section because this module is currently untested.
When I deleted and ran this again I had an issue related due to the project path. I changed the code to make it unique prior to creation of the moseq model
View entire conversation on ReviewNB |
Description
Implementation of moseq withing spyglass pipelines.
PositionOutput
PositionOutput.get_pose_dataframe
to allow fetching of full pose rather than centroid for appropriate pipelines.PoseGroup
used to combine multiple epochs of pose data into a training set (analagous toPositionGroup
in decoding pipeline)MoseqModelParams
MoseqModel
for reuse and extended trainingMoseqModelSelection
combinesPoseGroup
andmodel parametersMoseqModel
trains model and stores in external project fileMoseqSyllable
(and it's selction table) apply a trained model to an entry fromPositionOutput
AnalysisNwbfile
entryAdded utility function
get_position_interval_epoch
My To-dos
PositionOutput
Checklist:
CITATION.cff
alter
snippet for release notes.CHANGELOG.md
with PR number and description.