Set HyperTransformer configuration manually #995
Conversation
Codecov Report
@@ Coverage Diff @@
## master #995 +/- ##
==========================================
+ Coverage 70.24% 70.46% +0.21%
==========================================
Files 38 38
Lines 2850 2878 +28
==========================================
+ Hits 2002 2028 +26
- Misses 848 850 +2
Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here. |
| for name, dtype in dtypes.items(): | ||
| dtype = np.dtype(dtype).kind | ||
| field_metadata = self._fields_metadata.get(name, {}) | ||
| transformer_template = field_metadata.get( |
There was a problem hiding this comment.
does this take the metadata into account at all? I'm trying to figure out why it wasn't before and what has changed now. For example in the student_placements demo it wasn't making the duration column categorical even though the metadata said it was
There was a problem hiding this comment.
Old code was reading properly the dtype but was printing the wrong config which leaded to confusions as the transformers were updated after the detect. Now both sdtypes and transformers are being set corresponding to the dtype from the metadata with the set_config method.
00c9a77 to
4dfad77
Compare
|
|
||
| if transformer_template is None: | ||
| sdtypes[name] = self._DTYPES_TO_TYPES.get(dtype, {}).get('type', 'categorical') |
There was a problem hiding this comment.
Is there a unit test that checks this case (when transformer_template is None)?
Resolves #982