29th January, 2020
Plans to attend: RyanW
Attending: RyanW, Josh, Jackson Maxfield Brown, Jamie, Ward,
-
Standing items
- Introductions: new joinees should feel welcome to introduce
themselves and their interest in zarr/n5.
- Jackson+Jamie: Allen Institute for Cell Science, deal with
“Zeiss” CZI files, have python bindings to “libczi”, integrating with Dask, interested in Zarr for storage.
- Ward: Unidata (Boulder, CO); team lead for netcdf-c
(netcdf-java folks v interested in Java-based Zarr impl)
- Ryan: Mount Sinai School of Medicine in NYC; scRNA-seq biofx
→ Zarr
- John: HHMI→Nvidia, lots of image data work → Zarr
- Introductions: new joinees should feel welcome to introduce
-
NumFOCUS / Quansight
-
discussing Quansight doing contract work related to EOSS grant
- James Bourbeau (dask) may be taking this on, but his time is limited
- But looking for someone to focus on the v3 spec in 2020
-
Zarr is now a NumFOCUS project!
-
-
Zarr v2: any open and/or blocked PRs?
-
Jamie: what features are being added?
- John: less adding things and more removing things. E.g. Zarr in
Python allows you to store objects
-
Josh: smaller core with an extension mechanism
- [*Zarr v3
update*](https://zarr-developers.github.io/zarr/specs/2019/06/19/zarr-v3-update.html) (June 2019)
-
-
Goal: have OME move a huge amount of archival data to Zarr
-
today: OME-TIFF; more features desired
- some issues keeping XML aligned w/ TIFFs today; possible to put
relevant metadata into Zarr, avoid extra level of metadata indirection around Zarr?
-
recently launched tool for converting out of proprietary formats
- Zarr/N5 produced as intermediates
- might be interesting to use those as 1st-class
-
-
NetCDF:
- merged 1st code from nc-zarr dev fork into main branch
- Ward hopefully speaking at [*EGU in Vienna in
May*](https://egu2020.eu/register_and_venue/registration.html)
- John: does it make sense for people writing bindings to look at
it before release?
- Ward: not sure without speaking to Dennis. Perhaps the next
month or two.
- merged 1st code from nc-zarr dev fork into main branch
-
Opportunities for contributions from Allen Institute folks?
-
interest in Zarr from C(++)
- see: z5 (“Lightweight
C++ and Python interface for datasets in zarr and N5 format”)
- [*zarr-python
issues*](https://github.com/zarr-developers/zarr-python/issues?q=is%3Aissue+is%3Aopen+sort%3Aupdated-desc)
-
Josh: bioformats2”raw” & metadata pain points
- Jamie: in house Python code to parse metadata.
- Jackson: hoping to have people writing Zarr ASAP. Dask clusters aren’t giving the desired speed up.
- https://github.com/allencellmodeling/aicsimageio
- Using it in production and migrating as necessary
-
-
AOB (Any Other Business)
- Josh: Shared gdoc folder. John: use NumFOCUS folder? RyanW: not
sure it’s shared. Start off by adding people to the current folder.
- [*Here’s a “community” gdrive
folder*](https://drive.google.com/drive/folders/144M9w9LQElyu7gMGeT5tpqGvibqZKCJe?usp=sharing) with this doc in it
- Josh: Shared gdoc folder. John: use NumFOCUS folder? RyanW: not