Open chrisbarber opened 6 years ago
These should be callable via a very similar interface to the correlation search. E.g. domain1 and domain2 should be selectable via a set of select, coords, filters.
https://github.com/AllenInstitute/datacube/blob/dev/services/pandas/datacube/differential.py https://github.com/AllenInstitute/datacube/blob/dev/services/pandas/datacube/fold_change.py
(note: it looks like the _mask_args method didn't get copied over in https://github.com/AllenInstitute/datacube/issues/5)
_mask_args
This was partly done for differential search at one time on a demo branch:
http://stash.corp.alleninstitute.org/users/chrisba/repos/datacube/browse/services/pandas/datacube/datacube.py?at=refs%2Fheads%2Fbob-demo#724-732
The fold_change and differential implementations there look simpler-- it may be better to go with those instead:
http://stash.corp.alleninstitute.org/users/chrisba/repos/datacube/browse/services/pandas/datacube/datacube.py?at=refs%2Fheads%2Fbob-demo#77,95
Fold change and differential methods should be fitted with caching where possible as in the correlation search e.g.
https://github.com/AllenInstitute/datacube/blob/dev/services/pandas/datacube/datacube.py#L70
Also, correlation search should be moved to its own module for consistency with these.
Needed for #99 (differential is, at least)
These should be callable via a very similar interface to the correlation search. E.g. domain1 and domain2 should be selectable via a set of select, coords, filters.
https://github.com/AllenInstitute/datacube/blob/dev/services/pandas/datacube/differential.py https://github.com/AllenInstitute/datacube/blob/dev/services/pandas/datacube/fold_change.py
(note: it looks like the
_mask_args
method didn't get copied over in https://github.com/AllenInstitute/datacube/issues/5)This was partly done for differential search at one time on a demo branch:
http://stash.corp.alleninstitute.org/users/chrisba/repos/datacube/browse/services/pandas/datacube/datacube.py?at=refs%2Fheads%2Fbob-demo#724-732
The fold_change and differential implementations there look simpler-- it may be better to go with those instead:
http://stash.corp.alleninstitute.org/users/chrisba/repos/datacube/browse/services/pandas/datacube/datacube.py?at=refs%2Fheads%2Fbob-demo#77,95
Fold change and differential methods should be fitted with caching where possible as in the correlation search e.g.
https://github.com/AllenInstitute/datacube/blob/dev/services/pandas/datacube/datacube.py#L70
Also, correlation search should be moved to its own module for consistency with these.