Jake Vanderplas

That post predates newer functionality; IPython 7.3 introduced pip and conda magic functions so that ``%conda install package`` essentially does what you want it to, without having to put too...

I think there are two different modes of work: interactive exploration & data analysis (which suits itself well to Jupyter notebooks and packages installed within the active kernel on the...

I think the slowness comes from test patterns that look like this: https://github.com/data-apis/array-api-tests/blob/f82c7bc8627cc2c3a44fa3e425f53a253a609aa8/array_api_tests/test_creation_functions.py#L358-L364 In JAX, each operation outside the context of JIT compilation has a small amount of overhead related...

I pasted the slowest 20 tests below. For the most part I think it comes down to the slow repeated indexing I mentioned above: every one of these that I've...

Would you be open to changing logic that looks like this: https://github.com/data-apis/array-api-tests/blob/f82c7bc8627cc2c3a44fa3e425f53a253a609aa8/array_api_tests/test_creation_functions.py#L358-L364 to something more like this? ```python k = kw.get("k", 0) expected = [[1 if j - i ==...

Another option would be to use dlpack to export the full array to `numpy` or some other standard, where eager-mode indexing is not problematic.

Thanks - with `--max-examples=5` and a number of skips related to mutation and other issues, the `jax.experimental.array_api` PR now passes the test suite! https://github.com/google/jax/pull/16099 I did find that it errors...

Thanks – your chart contains no direct data transforms, so the code lets it pass through unmodified. There are transforms implied by ``mark_errorbar``, but altair-transform does not yet recognize those....

Some of these transforms have been implemented on master, but there hasn't been a release of this package for a while/

Sounds good! I'm glad to hear you're working on that. Have you tackled evaluation of vega expressions yet? That was probably >50% of the effort that went into this repo,...