Add skeleton of meta and dataset modules. #1
Conversation
nimare/meta/ibma/ibma.py
Outdated
| from ..base import MetaEstimator | ||
|
|
||
|
|
||
| class MFX_GLM(MetaEstimator): |
There was a problem hiding this comment.
Will probably inherit from IBMAEstimator
nimare/meta/ibma/ibma.py
Outdated
| The gold standard image-based meta-analytic test. Uses contrast and standard | ||
| error images. | ||
| """ | ||
| def __init__(self, dataset, n_iters=10000, voxel_thresh=0.001, clust_thresh=0.05, |
There was a problem hiding this comment.
I think we'll almost certainly need another layer on top of dataset that represents the actual component of a database/dataset that can be submitted for analysis. So I still lean strongly towards making dataset an argument of fit and not init. Otherwise we'll end up in a weird situation where both the Dataset and the MetaEstimator classes are implementing methods that filter/mutate the data.
nimare/dataset/dataset.py
Outdated
| except: | ||
| raise Exception('Nope') | ||
|
|
||
| def get(self, search='', algorithm=None): |
There was a problem hiding this comment.
Given the scope of the project and the variety of different data types this object is going to store, I think we may want more explicit accessor methods (e.g., get_studies, get_metadata, etc.). Fine to leave this as a placeholder, but ultimately we'll probably replace it with a bunch of more specific methods.
There was a problem hiding this comment.
I'll add in some more empty methods, just so there's a blueprint.
nimare/meta/cbma/mkda.py
Outdated
| from ...utils import intersection, diff | ||
|
|
||
|
|
||
| class MKDA(MetaEstimator): |
There was a problem hiding this comment.
Should inherit from CBMAEstimator
- Also add MetaResult class, and parent classes for contrasts and data types.
|
I don't think I can push commits to your repo, so I'll merge as-is and then create a new branch with edits. |
Also, add ALE and SCALE code from pyale. It won't work though.