Skip to main content Link Menu Expand (external link) Document Search Copy Copied

2023-03-22

Attending: Hailiang Zhang (HZ), Sanket Verma (SV), Dieu My Nguyen (DMN), Dennis Heimbigner (DH), John Kirkham (JK), Norman Rzepka (NR), Jeremy Maitin-Shepard (JMS), Johana Chazaro-Haraksin (JCH), Davis Bennett (DB)

TL;DR:

SV started the community meeting with a few updates. One of the most important updates was that we have a new page on the website for the Zarr adopters. Check here: https://zarr.dev/adopters/.

After this, HZ started presenting his ZEP, post which there was a Q&A session. After the Q&A session, we concluded the meeting with a few action items for HZ, which he’ll take care of in the upcoming weeks.

Updates:

Meeting minutes:

  • Hailiang’s ZEP0005 presentation - check the recording here! 🎥

  • Q&As post the session are below 👇🏻:

    • JMS: Mathematical equation would be helpful
      • HZ: Working on a paper for more details - internal for now - will be publishing it soon!
    • DB: Is it similar to summed area table? - https://en.wikipedia.org/wiki/Summed-area_table
      • HZ: Not exactly - Trying to achieve something more centric to Zarr - making the accumulation flexible and dimension agnostic
    • JMS: I think your is trying to solve the same problem as a summed-area table solves. But I think you’re trying to do it without requiring to storing something the same size as the original array - but perhaps imposing some additional restrictions on the type of queries you can do
      • HZ: This is more Zarr related - Jeremy has more comments on the PR - stride boundary aligned with the Zarr boundary
    • JMS: Don’t understand it fully - a mathematical equation would be helpful
    • SV: Is there a reference implementation?
      • HZ: I’m working on the code - will open-source it by the end of the summer
    • HZ: What is not clear, Jeremy?
      • JMS: Does the chunk needs to be aligned?
      • HZ: For this implementation the chunks need to be aligned
    • JMS: Trying to understand the proposal in a general sense
      • HZ: We can achieve the accumulation/averaging service faster - Giovanni will be using this accumulation
    • DB: Would you share some cloud backed example?
      • HZ: Sure, can do!
    • SV: Having a .zarr data before and after the accumulation chunks with the attribute would be good for everyone!
      • HZ: Sure, can do and can also add the script for the data generation!
    • DB: Having a reference of summed area table would be good thing!
      • HZ: Sure, can do that!