Implement dataless cubes #6253
Implement dataless cubes #6253
Conversation
for more information, see https://pre-commit.ci
for more information, see https://pre-commit.ci
for more information, see https://pre-commit.ci
for more information, see https://pre-commit.ci
Codecov ReportAttention: Patch coverage is
Additional details and impacted files@@ Coverage Diff @@
## main #6253 +/- ##
==========================================
- Coverage 89.83% 89.74% -0.10%
==========================================
Files 88 88
Lines 23347 23467 +120
Branches 4344 4383 +39
==========================================
+ Hits 20974 21060 +86
- Misses 1646 1664 +18
- Partials 727 743 +16 ☔ View full report in Codecov by Sentry. |
This reverts commit 6ed270d.
|
Something is wrong with the _assert_axiom code now. I need to catch a bus, so will have to figure out why tomorrow. |
for more information, see https://pre-commit.ci
|
Although the tests are passing, next time I look at this, I need to investigate the data setter. Currently, line 256 causes issues, as self._shape should be changable if self._shape = (), and it won't be. That is to say, it works as expected, but I believe it hasn't been written in such a way as it SHOULD be. |
lib/iris/_data_manager.py
Outdated
|
|
||
| @property | ||
| def shape(self): | ||
| """The shape of the data being managed.""" | ||
| return self.core_data().shape | ||
| return self._shape if self._shape else self.core_data().shape |
There was a problem hiding this comment.
This is defensive; it wouldn't cause any issues that I can think of to just have return self._shape, but this should protect us in case I've missed something.
There was a problem hiding this comment.
I've actually removed this now.
() is falsy. I could have easily done ...if self._shape is not None, but I decided there's no point in adding redundant code. Easy undo if my reviewer disagrees!
| state = is_lazy ^ is_real | ||
| assert state, emsg.format("" if is_lazy else "no ", "" if is_real else "no ") | ||
|
|
||
| if not (is_lazy ^ is_real): |
There was a problem hiding this comment.
This might be clearer as if is_lazy == is_real.
There was a problem hiding this comment.
I did see this, but forgot to comment.
Personally, I think I prefer it as it is, but if there's significant desire for your suggestion, I don't mind.
Closes #4447.
I plan to do this in four stages:
data is Nonewhen handed a shape value, and can essentially round-trip removing and adding data. This shouldn't break existing tests. 71c7ae8DataManagermethods all make sense and work withNonedata.DataManager.copy()is an example of a method that won't make sense withNonedata.