-
-
Notifications
You must be signed in to change notification settings - Fork 364
remove chunk encoding from config #3228
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Conversation
Codecov Report✅ All modified and coverable lines are covered by tests. Additional details and impacted files@@ Coverage Diff @@
## main #3228 +/- ##
==========================================
- Coverage 94.76% 94.76% -0.01%
==========================================
Files 78 78
Lines 8642 8672 +30
==========================================
+ Hits 8190 8218 +28
- Misses 452 454 +2
🚀 New features to boost your workflow:
|
closes #3104 |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I think that makes sense!
FYI @d-v-b there's still some references to |
Are these not fixed by #3390? |
Possibly! I'm behind on reviews :) |
This PR removes the default chunk encoding parameters from the config.
Prior to the data types refactor, when we inferred the chunk encoding automatically, we categorized the array data type into one of a few categories like "numeric", "string", "bytes". These categories were created before we added datetimes and fixed-length strings, and they don't really make sense after those data types were added:
It's not really possible to preserve the old categories in the config now that we have more data types that resist categorization. I don't think it's really possible break the data types down into a small number of general categories, and it's less possible / attractive to put this categorization logic in our config.
I think a better solution is for zarr python to choose sane defaults; when users want something other than the default, they specify it in the function call. In practice these sane defaults are the same codecs for all data types except the variable-length data types, which get special treatment (data types with endianness in zarr v3 also get a
bytes
codec with an endianness parameter). That's what I added in this PR. The default chunk encoding parameters each come from functions that are inaccessible to configuration.I did a quick github survey to find users who might be using the zarr config to change the default codecs, and I could only find 2: @TomAugspurger uses the config here and @dcangst in here.