Tightened semantics of Dimension objects #1245
Conversation
holoviews/core/dimension.py
Outdated
| raise Exception("Values argument can only be set with the string 'initial'.") | ||
| values = params.get('values', []) | ||
| if isinstance(values, basestring) and values == 'initial': | ||
| self.warning("The 'initial' string for dimension values is no longer supported.") |
There was a problem hiding this comment.
Unless I'm confused set values to an empty list here, otherwise param will throw an exception after you've already warned.
There was a problem hiding this comment.
Actually no, you'll get ['a', 'i', 'i', 'i', 'l', 'n', 't'].
There was a problem hiding this comment.
Fixed in 9d57a58 (I pushed forced an amended commit)
jlstevens
force-pushed
the
dimension_cleanup
branch
from
87a4113 to
9d57a58
Compare
|
Looks good to me and I'm happy with the pprint. |
|
Great! I'm happy the tests have passed - I was worried there might be double specification of I am also considering making |
jlstevens
force-pushed
the
dimension_cleanup
branch
from
5bbb9f5 to
21443cf
Compare
|
Rebasing against current master to get the push builds passing (pr builds are green). |
jlstevens
force-pushed
the
dimension_cleanup
branch
from
b7640d4 to
24a0e88
Compare
|
The pr build has now passed for Python 3 and 3. I will now try to eliminate old uses of Note, the next commit is one that I'll use to find where dimensions are cast to strings. It uses some nasty inspection code and I'll keep amending and push forcing that commit till I've caught them all. |
|
Rebasing again due to recent notebook test data updates. |
jlstevens
force-pushed
the
dimension_cleanup
branch
5 times, most recently
from
59e4bcf to
2fd9fe3
Compare
jlstevens
force-pushed
the
dimension_cleanup
branch
from
2fd9fe3 to
e503ec4
Compare
holoviews/core/dimension.py
Outdated
| @@ -64,7 +63,8 @@ def replace_dimensions(dimensions, overrides): | |||
| elif isinstance(override, Dimension): | |||
| replaced.append(override) | |||
| elif isinstance(override, dict): | |||
| replaced.append(d(**override)) | |||
| replaced.append(d(override.get('name',None), | |||
There was a problem hiding this comment.
Should be using .clone internally now?
There was a problem hiding this comment.
Yes, __call__ is now an alias to clone but I do want to recommend clone from now on. Thanks for point it out and I'll fix it shortly.
holoviews/core/dimension.py
Outdated
| elif (name,) in self.presets.keys(): | ||
| existing_params = dict(self.presets[(str(name),)].get_param_values()) | ||
| elif spec in self.presets.keys(): | ||
| existing_params = dict(self.presets[str(spec)].get_param_values()) |
There was a problem hiding this comment.
This doesn't seem right, you check spec is in presets but then do the lookup with str(spec)? Same below, also is .keys() needed in the conditional?
There was a problem hiding this comment.
Addressed in 703c01d
|
Tests have passed and I think this PR is now ready to review. |
holoviews/core/dimension.py
Outdated
| if name is not None: settings['name'] = name | ||
| return self.__class__(**settings) | ||
| spec = spec if (spec is not None) else (overrides.get('name', self.name), | ||
| overrides.get('label', self.label)) |
There was a problem hiding this comment.
Don't want to check if name or label are supplied along with a spec?
There was a problem hiding this comment.
Does it matter? If there is a spec, it gets passed through as the first argument, shouldn't matter if it is just a string or just a tuple.
There was a problem hiding this comment.
Hmm, since you're filtering out name and label from kwargs below, those will simply be silently be dropped.
There was a problem hiding this comment.
Ah got you. Fixed in 6d50a8a.
|
@philippjfr I've addressed your comments and the tests are passing. Any other suggestions for dimension improvements? |
|
No, should I merge? |
|
If you are happy with the changes, please do! |
|
This pull request has been automatically locked since there has not been any recent activity after it was closed. Please open a new issue for related bugs. |
This PR aims to address issue #843 and clean-up the definition of Dimension objects generally.
It is still work in progress and I've started this PR at the first commit just so that I can see the results of the notebook tests as I progress.
Edit: I have copied over the action items from the issue:
nametospec.labelas a parameterlabelto be passed as a kwarg and warning when ambiguous.__call__method, making it work with(name, label)tuples__call__toclone.specproperty