Simon Kornblith
Simon Kornblith
Moving discussion about how to represent missing values in DataArrays from JuliaLang/julia#9363 and JuliaStats/DataFrames.jl#745 here. I've done some benchmarks [here](https://gist.github.com/simonster/b1b4cc2ad0daa8e20a99) that show that, with the best code I can generate...
At the moment, `setindex!` has to look up every value in the pool, so the computational complexity of setting `n` indices in a PooledDataArray with a pool of size `m`...
It'd be nice to have a subtype of AbstractDataArray that wraps a regular floating point array and treats `NaN` values as `NA`. I think this should generally be faster than...
With 440c9bcdf2204d85e6dc31231eb0c8208756fd1b: ``` julia> mean([1, 2], 1) ERROR: InexactError() in scale! at linalg/generic.jl:16 julia> @which mean([1, 2], 1) mean(::DenseArray{T
At present, [this check](https://github.com/google/jax/blob/7a3e2140fbda4674da855f611f515d271adba1a1/jax/_src/compiler.py#L387) stops the compilation cache from being written on any process that is not process index 0. This makes sense if the compilation cache directory resides on...
They're not exported, but since we refer to them in deprecation warnings, they should probably be documented.
The description says that there are 50000 images. However, it seems like there are 50889. Some synsets contain sketches numbered 0-50 whereas others contain sketches numbered 0-49.
The DICOM files I get from our PACS here at MIT Martinos Center don't have headers. Is this common enough that `dcm_parse` should just try to read the file anyway...
This requires some additional state while parsing.
If the content of the `(0028, 0103)` tag is 1, then the pixel data is signed, but at present `pixeldata_parse` always treats data as unsigned.