-
Notifications
You must be signed in to change notification settings - Fork 125
Open
Conversation
Adds: - `read_zarr` and `write_zarr` methods which fall back to pickle if no zarr store is available. - Warnings that read_pickle methods will be deprecated for future GlacierDirectories. - Zarr as a core dependency. Refs: OGGM#1903
Adds: - Validation for different zarr/pickle stores to maintain backwards compatibility. - Dedicated module for zarr-related operations. - Conversion between existing pickles and zarr. - Reconstruction of OGGM objects from minimum possible data. Refactors: - Replaces all instances of `read_pickle` with `read_store`. Compatibility is maintained
Currently writes each data set to a separate file instead of a larger data store
ON-60 write data to zarr
Falls back to pickle if zarr r/w fails Refs: ON-60
... into ON-60-write-data-to-zarr
also removes hardcoded pickles
Merge fix pickle on 60 on 63
ON-60 write data to zarr
fix(workflow): replacement zarr groups
It's finally here :)
In existing documentation the convention is to explicitly add a separator.
WARNING: This branch includes ON-94 and ON-55. Adds: - New file format for storing glacier directories (v2). - Framework to stream glacier directories from zip files. Backwards-compatible. - Utilities for batch converting existing v1 gdirs to v2 and vice versa. - Mock streaming utility until v2 gdirs are available on the cluster.
Refactors ffspec chain with a RemoteBundleStore for better cache integration. Adds support for reading gdirs remotely.
@gampnico
gampnico
changed the title
(削除) Feat on 95 stream glacier directories (削除ここまで)
(追記) feat: streamable glacier directories (追記ここまで)
Jul 30, 2026
...ream-glacier-directories
@gampnico
gampnico
changed the title
(削除) feat: streamable glacier directories (削除ここまで)
(追記) feat: stream glacier directories (追記ここまで)
Aug 10, 2026
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Uh oh!
There was an error while loading. Please reload this page.
CAUTION: This branches off from #1908, #1951 for zarr support and incremental glacier directory creation. Those should be merged first.
Closes #1896
For clarity, file formats are:
legacy: the cumulative tar system currently used by OGGM v1.6
v1: the incremental tar system introduced by #1951
v2: the streaming zip system introduced by this PR.
Each format is backwards-compatible with the previous ones.
For a full overview of v2 system and its capabilities I strongly recommend running the notebook here .
In concrete terms, users can now:
Technical Overview
Streaming-format glacier directories store array data (gridded, climate) and the former pickle objects (flowlines, geometries, etc) as zarr groups inside a single per-glacier
data_store.zarrstore, and vector data as GeoParquet.The user-facing API is unchanged:
workflow.init_glacier_directories(..., from_prepro_level=n, prepro_base_url=...). The client probes for.zipbundles first and falls back to the classic.tarbundles transparently, so old base URLs and old datasets keep working.utils._downloads.peek_remote_manifestandutils._downloads.open_remote_groupread a manifest or lazily open a zarr group out of a remote bundle via HTTP range requests so no downloading to disk.Existing published datasets can be converted with two functions from
oggm.utils.compat:convert_prepro_to_deltas(..., convert_to_zarr=True)(legacy to incremental v1, pickles to zarr) andconvert_prepro_to_v2_artifacts(v1 to v2 streaming artifacts).Tests added/passed
Fully documented
Entry in
whats-new.rst