Lachlan Deakin
Lachlan Deakin
> @normanrz @dstansby @LDeakin @zoj613 thoughts on the changes proposed in this PR? In particular the statement that implementations SHOULD support the core codecs, but also anything else. It all...
> worth noting that this changes some defaults (we were using a default fill_value of `None` in some places, and 0 in other places, for example). When in doubt, I...
> how did you check this? because for zarr v2 data, it _should_ make a difference -- `None` is a valid v2 fill value, and it's not the same as...
> I think for v2 data the contents of the array metadata will be sensitive to the None / 0 distinction, but at runtime the actual array data should be...
Codecs that do pass-through decoding like `bitround`, `quantize`, etc., could declare this. Viewers can read such arrays without explicitly adding support for the codecs.
I agree that codecs should not be ignored when writing data. This is similar to what I was arguing for with `extensions`: https://github.com/zarr-developers/zeps/pull/67#issuecomment-2907501279 > `must_understand: false` in an extension implies...
`zarrs` supports all of them for reading. https://docs.rs/zarrs_metadata/latest/zarrs_metadata/v3/struct.MetadataV3.html#example-metadata. However, codecs are always serialised using variant 2-5 since https://github.com/zarrs/zarrs/commit/acaad11ce0dbb5d5446fe7f12c969791abe82001. > ### Fixed > Fixed a regression introduced in v0.11.2 ([89fc63f](https://github.com/LDeakin/zarrs/commit/89fc63fa318cfd780e85fec6f9506ca65336a2c3)) where...
There is also another variant since Zarr 3.1: `{"name": "bytes", "configuration": {}, "must_understand": true}`. I technically support this in `zarrs` and would like to start writing `"must_understand": false` with a...
I'm fairly certain this was discussed in an issue/PR on GitHub, but I can't find it. Maybe @normanrz mentioned it? The spec permits it: > An extension object MAY explicitly...
I don't think it should be reverted. It is a sensible shorthand and it is consistent with other array metadata fields. I think it is a valid change given the...