Add version limits to dependencies. #304
Conversation
| "numpy~=1.15", | ||
| "numpy>=1.15.0,<1.21.0", |
There was a problem hiding this comment.
I don't think we need to pin numpy here since TensorFlow tends to have a constaint on the numpy version used anyway.
There was a problem hiding this comment.
I'm not sure, I think we do need to declare it as a dependency because it is used in our code. Like, if somebody installed Zoo in a fresh environment, without Numpy in our deps they might see an import error for Numpy, when what they actually are missing is TF.
There was a problem hiding this comment.
Oh yeah, I just meant we can probably keep it as numpy~=1.15 here, but I am also OK with limiting it to numpy>=1.15.0,<1.21.0 if you prefer.
There was a problem hiding this comment.
Yeah I'd probably rather stick with this if that's okay, just for consistency if nothing else.
@lgeiger I think this should prevent what happened with #303, and Dependabot will bump the version ceilings for us. But it means we'll know if new versions break things.