Add to_dataframe to InferenceData #1395
Conversation
ahartikainen
changed the title
Codecov Report
@@ Coverage Diff @@
## master #1395 +/- ##
==========================================
+ Coverage 91.63% 91.65% +0.01%
==========================================
Files 105 105
Lines 10965 11022 +57
==========================================
+ Hits 10048 10102 +54
- Misses 917 920 +3
Continue to review full report at Codecov.
|
ahartikainen
force-pushed
the
feature/to_dataframe
branch
from
8d57933
to
08c22d8
Compare
| def __getitem__(self, key): | ||
| """Get item by key.""" | ||
| if key not in self._groups_all: | ||
| raise KeyError(key) | ||
| return getattr(self, key) |
| ) | ||
|
|
||
| if not group_names: | ||
| raise TypeError("No valid groups found: {}".format(groups)) |
There was a problem hiding this comment.
Is this logic ok?
posterior + observed_data --> df from posterior
observed_data -> TypeError
There was a problem hiding this comment.
I think it is ok, given that it fits our goals and we document it. My only question not really knowing yet how panel works is if ppc checks will be possible with it if observed_data is not in the dataframe
There was a problem hiding this comment.
Panel works with xarray too. My gist was just short cutting some code with data frame.
Description
Add
.to_dataframefunction to InferenceData. The result is a dataframe in a "wide" format.This PR also adds
__getitem__magic for the InferenceData.Checklist