Daniel Gafni

I generate the partitions from a yaml file. The file is checked into the repo and can be updated on-demand by executing one of the Python files, but the updating...

The issue is fixed for me. My offsets are the same tho. P.S. it seems like it would be better if `PartitionsDefinition` could be automatically sorted without explicitly defining sorting...

Would be very happy if this gets merged! I'm experiencing the same issue. @garethbrickman have you guys noticed this PR?

In general, applying `stylix` should be more versatile. I think it would be very useful if: - it was possible to disable some of (fonts, colors, ...) - it was...

Not sure if this is the correct place to write this, but... For the native partitioned Parquet reader, would it be possible to support loading unions of columns from different...

I've encountered this issue with `ArgoCD` and `argocd-image-updater` - image updater creates ArgoCD Helm parameters like `deployments[0].image.tag = my-tag`, which overwrites the whole `deployments[0]` map.

I tried implementing the required changes in `helm/dagster/charts/dagster-user-deployments/templates/helpers/_helpers.tpl`, but failed miserably. I tried to write a `map` -> `array` deployments converter, something like this: ```tpl # _helpers.tpl {{- define "dagsterUserDeployments.deploymentsArray"...

Hey @leycec , thanks for the explanation! I had a false assumption I could kinda keep the type annotations framework-agnostic and reuse the same function signature for both `numpy` and...

Oh amazing! Would `TypeVar` also work? This way I could check if a function inputs and outputs are either all numpy or all torch, for example.

Seems like it doesn't: ```python from typing import TypeVar, Union import numpy as np import jax.numpy as jnp from jaxtyping import Float AnyArray = TypeVar( "AnyArray", bound=Union[np.ndarray, jnp.ndarray], ) Float[AnyArray,...