why do boolean comparisons with Py object return Py object?

[schlichtanders]

First time I am trying to use pythoncall support for defining python classes. Apparently all input arguments as well as self attributes get wrapped in Py.

This is also truth for boolean comparisons like == or != Which return Py objects.

This is difficult as we cannot simply use standard comparisons in if else && or ||, but instead need to wrap everything in Bool(...) In order to get a Julia boolean

What is the purpose of this extra complication? Wouldn't it be much easier if such boolean operators return Julia Bool values?

[cjdoris]

This is because Python boolean operators do not always return booleans. For example, comparing two numpy arrays will return an array of booleans.

Some options:

• Call pytruth (or Bool) as you already do.
• Call pyeq(Bool, x, y), which is equivalent to pytruth(x == y) (and similar for other comparisons).
• Call isequal(x, y), which is equivalent to the above too.

[schlichtanders]

Thank you for the explanation. Would it be possible to nevertheless return a Julia Bool if a standard Boolean was returned?
I think this would massively simplifiy writing Py classes from Julia (for the initially stated reasons)

[cjdoris]

I'm not going to do that I'm afraid, as it introduces type instability which is not only bad for performance but also makes code harder to reason about. This is one of the major design differences between this and PyCall (which automatically converts everything).

[github-actions]

This issue has been marked as stale because it has been open for 30 days with no activity. If the issue is still relevant then please leave a comment, or else it will be closed in 7 days.

[github-actions]

This issue has been closed because it has been stale for 7 days. You can re-open it if it is still relevant.