TEST: include `icechunk`s zarr tree strategies here by ianhi · Pull Request #3914 · zarr-developers/zarr-python

ianhi · 2026-04-16T20:53:52Z

Up streaming our zarr tree hypothesis strategy from icechunk.

generates trees like these:

/
├── EtTWAYY/
│   ├── EtTWAYY.jEm  [array shape=(1,)]
│   ├── EtTWAYY/
│   │   ├── 7ek  [array shape=(1,)]
│   │   ├── g  [array shape=(1,)]
│   │   ├── 7H7_NaN  [array shape=(1,)]
│   │   └── NaN  [array shape=(1,)]
│   ├── NaN  [array shape=(1,)]
│   └── 91  [array shape=(1,)]
└── NaN/
    └── srtB  [array shape=(1,)]

Two commits

398ecb4 adds the models and strategies - nearly a straight lift from icehcunk. added a bit more logic around materialize write mode, and zarr format + case sensitivity otherwise the same as in icechunk.
bb62db7 uses them in some tests - mostly as an example for anyone who wants to add more in the future.

Add unit tests and/or doctests in docstrings
Add docstrings and API docs for any new/modified user-facing classes and functions
New/modified features documented in docs/user-guide/*.md
Changes documented as a new file in changes/
GitHub Actions have all passed
Test coverage is 100% (Codecov passes)

…trategy Introduces declarative tree descriptors in `zarr.testing.models` (ArrayNode, GroupNode) with `materialize()` / `from_store()` round-trip helpers, and a hypothesis `trees()` strategy in `zarr.testing.strategies` that generates realistic zarr hierarchies with prefix-colliding names (sibling `foo`/`foo-bar`, cousin name reuse). `trees()` accepts `case_insensitive: bool | None` to keep sibling names from differing only in letter case when targeting stores backed by a case-insensitive filesystem (macOS APFS, Windows NTFS default). Defaults to the current platform's filesystem behavior. `materialize()` accepts an optional `zarr_format` to exercise v2 and v3 from the same tree descriptor.

@example

Two property tests exercising zarr hierarchy code paths against the new trees() hypothesis strategy: - test_groups_and_arrays_split_children — verifies that Group.groups() and Group.arrays() return exactly the sub-groups and arrays respectively, across random hierarchies with prefix-colliding sibling names. Uses a MemoryStore since the partition logic is store-agnostic; local/zip coverage is already carried by the sibling imperative test. - test_group_members_tree_roundtrip — materialize() / from_store() round-trip for any GroupNode tree. Each test pins a hand-picked shape via @example so the known-good case runs alongside the random draws.

codecov · 2026-04-16T21:06:52Z

Codecov Report

❌ Patch coverage is 81.67939% with 24 lines in your changes missing coverage. Please review.
✅ Project coverage is 92.97%. Comparing base (dd5a321) to head (a8a2ca0).

Files with missing lines	Patch %	Lines
src/zarr/testing/models.py	72.50%	22 Missing ⚠️
src/zarr/testing/strategies.py	96.07%	2 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #3914      +/-   ##
==========================================
- Coverage   93.10%   92.97%   -0.14%     
==========================================
  Files          85       86       +1     
  Lines       11193    11324     +131     
==========================================
+ Hits        10421    10528     +107     
- Misses        772      796      +24

Files with missing lines	Coverage Δ
src/zarr/testing/strategies.py	`93.41% <96.07%> (+0.50%)`	⬆️
src/zarr/testing/models.py	`72.50% <72.50%> (ø)`

🚀 New features to boost your workflow:

❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

d-v-b · 2026-04-17T06:28:51Z

+class GroupNode:
+    children: dict[str, ArrayNode | GroupNode] = field(default_factory=dict)
+
+    def walk(self, prefix: str = "") -> Iterator[tuple[str, Node]]:


i feel like "iter_nodes"or similar is a bit more expressive than "walk"

d-v-b · 2026-04-17T06:30:26Z

Thanks for this! I feel like if these array / group hierarchy methods are useful for testing, then they are just useful, and we should consider integrating them into the codebase. With that in mind I'm gonna leave some comments but they are just advisory, feel free to ignore.

d-v-b · 2026-04-17T06:32:31Z

+            if isinstance(child, GroupNode):
+                yield from child.walk(p)
+
+    def nodes(self, prefix: str = "", *, include_root: bool = False) -> list[str]:


all these iteration routines can do IO. so returning list[stuff] requires doing all the IO before the first value is available. Iterator[stuff] or Generator[stuff] gives us a bit more flexibility if we want to stream results instead of doing all the IO up-front.

(this applies in a world where we put these routines on the group class that's bound to a store)

d-v-b · 2026-04-17T06:36:18Z

+
+
+@dataclass(frozen=True)
+class GroupNode:


this approach to "structurally model a group" reminds me of GroupSpec in pydantic-zarr. this is clearly a useful representation, so we should figure out how we can fit it into zarr-python proper

d-v-b · 2026-04-17T06:38:04Z

+
+            GroupNode.from_store(some_memory_store)
+        """
+        return sync(cls.from_store_async(store))


using sync here forces any consumer to use the zarr-python event loop. flagging this for visibility

d-v-b · 2026-04-17T06:40:08Z

+    zarr_format=st.sampled_from([2, 3]),
+)
+@pytest.mark.filterwarnings("ignore:Consolidated metadata:zarr.errors.ZarrUserWarning")
+def test_groups_and_arrays_split_children(


was there bugged behavior that this test was necessary to fix?

ianhi added 2 commits April 16, 2026 16:49

github-actions bot added the needs release notes Automatically applied to PRs which haven't added release notes label Apr 16, 2026

changes

bffe644

github-actions bot removed the needs release notes Automatically applied to PRs which haven't added release notes label Apr 16, 2026

d-v-b reviewed Apr 17, 2026

View reviewed changes

Merge branch 'main' into ian/tree-strats

a8a2ca0

d-v-b reviewed Apr 17, 2026

View reviewed changes

d-v-b mentioned this pull request Apr 17, 2026

[v3] declarative hierarchy API #1754

Open

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

TEST: include `icechunk`s zarr tree strategies here#3914

TEST: include `icechunk`s zarr tree strategies here#3914
ianhi wants to merge 4 commits intozarr-developers:mainfrom
ianhi:ian/tree-strats

ianhi commented Apr 16, 2026 •

edited

Loading

Uh oh!

codecov bot commented Apr 16, 2026 •

edited

Loading

Uh oh!

d-v-b Apr 17, 2026

Uh oh!

d-v-b commented Apr 17, 2026

Uh oh!

d-v-b Apr 17, 2026

Uh oh!

d-v-b Apr 17, 2026

Uh oh!

d-v-b Apr 17, 2026

Uh oh!

d-v-b Apr 17, 2026

Uh oh!

d-v-b Apr 17, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

Uh oh!

Conversation

ianhi commented Apr 16, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

codecov bot commented Apr 16, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Codecov Report

Uh oh!

d-v-b Apr 17, 2026

Choose a reason for hiding this comment

Uh oh!

d-v-b commented Apr 17, 2026

Uh oh!

d-v-b Apr 17, 2026

Choose a reason for hiding this comment

Uh oh!

d-v-b Apr 17, 2026

Choose a reason for hiding this comment

Uh oh!

d-v-b Apr 17, 2026

Choose a reason for hiding this comment

Uh oh!

d-v-b Apr 17, 2026

Choose a reason for hiding this comment

Uh oh!

d-v-b Apr 17, 2026

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

ianhi commented Apr 16, 2026 •

edited

Loading

codecov bot commented Apr 16, 2026 •

edited

Loading