bionic
bionic copied to clipboard
square
A Python framework for data science.
Currently the helper function `bionic.util.init_basic_logging` configures the global logging state, with a default level of `INFO` for all loggers. However, I suspect most users only use this because they want...
When an already-existing entity is defined with a function, its previous declaration (including any protocol, docstring, etc.) is erased. This is intended behavior, but it can be surprising -- especially...
Entities marked with `@changes_per_run` generate a new cache entry on disk each time they're computed, even though these cached entries are unlikely to ever be used again. We should write...
When we run `generate_test_compatibility_cache.py` to update our regression test vectors, it doesn't delete the old test vectors, but there's no point in keeping them around. It would be better if...
(This is a bit of an edge case, since it requires switching versioning modes. The only time I've encountered it was while trying to prepare a demo of the different...
Right now, Bionic users can see the progress of their computation by watching the printed logs. However, the stuff we want to log isn't exactly the same as the stuff...
Entity values in Bionic can be accessed in two ways: from another entity definition, or directly from a flow with `Flow.get`. Currently these two APIs offer different ways to transform...
Currently the only way to create multiplicity is by setting fixed values with `builder.assign` or `builder.set`. This means the number of values can't depend on other values in the system....
Right now, when you call `Flow.reloading()`, you get a new copy of the flow with reloaded state. I think it should also mutate the flow to use the new state....
Currently you can install Bionic via Pip, but not Conda. But some people want to use Conda! In some ways Conda would be nicer because it could also grab dependencies...