refactor: use functools.wrap in func_metadata decorator#347
Conversation
Signed-off-by: themattmorris <[email protected]>
|
Well that's a new error. I do see the line in I believe this is occurring because of this update to There is this PR out there for |
|
The pyspark fix has been made that was causing the installation failure. Would you be able to rerun the checks? |
|
@themattmorris seems like some issues that are specific to your branch |
β¦func_metadata having narrower types in signature Signed-off-by: themattmorris <[email protected]>
b0b03ec to
9be5a7e
Compare
β¦w that their source code has changed since being wrapped Signed-off-by: themattmorris <[email protected]>
|
So the scope of changes in this PR certainly ended up being more than I expected, but that was a side effect of having the function signatures being updated for functions decorated with Some of the changes were tests where import inspect
from sqlframe.standalone import functions as SF
print(inspect.getsource(SF.ascii))(clearly this is the source code of the inner function inside the decorator) |
eakmanrq
left a comment
There was a problem hiding this comment.
Thanks for the contribution @themattmorris!
fixes #342
I didn't see where
CALLING_CLASSwas being used and assumed perhaps it was something that was previously used? I also removed importfrom __future__ import annotationsbecause I didn't see where it was necessary in this module, but let me know if you disagree with either of these changes.Thanks for considering this submission!