Pretty `_repr_html_` of cube dimension metadata by soxofaan · Pull Request #955 · Open-EO/openeo-python-client

soxofaan · 2026-10-09T13:00:05Z

soxofaan · 2026-10-09T15:51:52Z

Current state of the visualization (just table for now):

note that the temporal and spatial extents are still the unfiltered extents of the original collection, instead of the actual cube extents

Copilot

🟡 Changes recommended

Temporal extent handling and HTML rendering contain runtime failures for supported metadata and open-ended intervals.

7 open findings

Allow metadata without a temporal dimension · New
Coerce dimension types before HTML escaping · New
Import dateutil.parser explicitly · New
Add tests for interval conversion and intersection · New
Allow None in temporal extent annotations · New
Define stable representation for non-concrete extents · New
Add VectorCube HTML representation tests · New

What changed in this PR

Adds richer Jupyter rendering for cube metadata, addressing #827, while tracking temporal extents through filtering operations.

Changes:

Adds HTML dimension tables to raster and vector cube representations.
Introduces temporal interval intersection and extent markers.
Adds rendering tests and a changelog entry.

File	Description
`CHANGELOG.md`	Documents the Jupyter representation.
`openeo/dates.py`	Adds datetime migration note.
`openeo/metadata/__init__.py`	Adds rendering and temporal filtering.
`openeo/metadata/_extents.py`	Defines special extent markers.
`openeo/metadata/_jupyter.py`	Renders dimension metadata tables.
`openeo/rest/_datacube.py`	Names process-graph HTML output.
`openeo/rest/datacube.py`	Combines metadata and graph rendering.
`openeo/rest/vectorcube.py`	Adds metadata rendering to vector cubes.
`openeo/util.py`	Adds datetime migration notes.
`openeo/utils/datetime.py`	Implements datetime intervals.
`tests/metadata/test_jupyter.py`	Tests table rendering.
`tests/metadata/test_metadata.py`	Tests metadata HTML output.
`tests/rest/datacube/test_datacube.py`	Tests cube HTML composition.

🧠 Review effort: Balanced

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.

Copilot

🟡 Changes recommended

Open-ended temporal intervals can raise errors, and the public TemporalDimension constructor is unintentionally breaking.

3 open findings

6 resolved since last review

🧠 Review effort: Balanced

Copilot

🟡 Changes recommended

Half-open temporal filters can currently produce a non-empty metadata extent when the query contains no matching instants.

5 open findings

2 resolved since last review

🧠 Review effort: Balanced

Copilot

🟡 Changes recommended

Dynamic temporal bounds can raise errors, and successive filters lose exclusive-end semantics.

1 open finding

Dynamic temporal bounds are incorrectly parsed as concrete dates · New

5 resolved since last review

🧠 Review effort: Balanced

…955

…nterval #955

Copilot

🟡 Changes recommended

Successive temporal filters lose end-exclusive semantics and can produce an incorrect non-empty extent.

1 open finding

Preserve end-exclusive semantics when serializing intersections · New

1 resolved since last review

Dynamic temporal bounds are incorrectly parsed as concrete dates

🧠 Review effort: Balanced

…955

related to #955, #827

Copilot

🔵 Needs a closer look

Public metadata API changes and cross-cutting extent propagation need final maintainer validation beyond the notebook-rendering feature.

No open issues 🎉

1 resolved since last review

Normalize CRS values before comparing bbox and dimension references

_{Review effort: Balanced}

soxofaan · 2026-10-26T10:08:00Z

I think this PR is ready as MVP for #827 (did enough of iterating together with copilot)
but I'm going to wait to merge because of CDSE/Terrascope code freeze at the moment (client side metadata handling leaks through in openeo-python-driver and openeo-geopyspark-driver)

soxofaan · 2026-10-28T10:39:17Z

screenshot of current state:

temporal dimension metadata is now more in line with specified temporal extent
spatial dimension metadata is marked as "non concrete" because given spatial extent is given in EPSG:4326, while original collection dimension metadata is in Auto42001, and current MVP does not yet implement simulation of how a backend would handle that (note for example that pyproj is not a required dep in client, so we need this fallback anyway)

soxofaan · 2026-10-28T13:42:03Z

@copilot can you collect and make an overview of all the "TODO" comments added/updated or handled in this PR?

Copilot · 2026-10-28T13:44:08Z

@copilot can you collect and make an overview of all the "TODO" comments added/updated or handled in this PR?

I reviewed the PR diff and grouped its TODOs below.

Outstanding TODOs added/clarified by this work

openeo/metadata/__init__.py: make spatial axis required instead of inferring it from the dimension name (currently kept optional for compatibility); normalize concrete spatial extents to NumericalInterval; consider exposing crs as reference_system; support more bbox inputs/geometries; validate/normalize temporal extents without breaking the driver-facing API.
Spatial filtering/resampling: implement antimeridian-crossing handling and CRS reprojection. In particular, SpatialDimensionsXY.filter_bbox currently falls back to non-concrete extents for mismatched CRSs, and resample_spatial / resample_cube_spatial retain bounds without reprojection when the CRS changes. This is why the screenshot shows non-concrete x/y extents for an EPSG:4326 filter on Auto42001 metadata.
Parameterized values: consider replacing the sentinel singleton with a class and moving it to a reusable location.
tests/metadata/test_metadata.py: add proper antimeridian detection/handling coverage.

TODOs removed/handled

The previous TODO to support spatial-dimension handling in CubeMetadata was addressed by adding x/y metadata and bbox filtering.
The old STAC-alignment TODO about supporting an x/y axis was addressed by adding SpatialDimension.axis; a follow-up TODO remains to require it rather than guess.
The old note about blindly selecting the last band/temporal dimension was removed. Multiple dimensions now trigger a warning and the first is selected; user-selectable choice is still not implemented.

Moved rather than new work

The #738 real-STAC-request-mocking TODO and the collection/item asset-handling TODOs appear relocated with test reorganization, not completed by this PR.
The #465 date-utility-location TODO appears in date utility code and is unrelated to the metadata behavior.

Copilot

🟢 Approval recommended

No active review issues remain for the richer Jupyter displays and metadata extent-handling changes.

No open issues 🎉

_{Review effort: Balanced}

soxofaan · 2026-10-28T14:54:58Z

iterated some more after reviewing with fresh mind after code freeze

illustration of current state:

improved since previous screenshot:

_NonConcreteExtent now has a reason, to make it easier to trace back where the non-concreteness comes from
with the COPERNICUS_30 example: spatial_extent is in same CRS as original dimension metadata, so extent of result cube is here concrete
temporal extent of the COPERNICUS_30 cube is outside the temporal extent of the collection, giving "empty" result: shown as _EmptyExtent

soxofaan · 2026-10-28T17:36:54Z

wanting to merge this PR, but maybe best to first create a release of current master

Trying to keep all this into account: - keep existing public API - support for parameterized/dynamic `filter_temporal` - handle empty temporal extent properly - calculate real temporal intersection - don't require metadata refs #827, #949

related to #955, #827

(re)group test functions in relevant test classes for better overview and easier future evolution refs #949 #827 #955

refs #827, #949

ref #827 #955

…cubes with z dimension' Co-authored-by: Copilot Autofix powered by AI <175728472+Copilot@users.noreply.github.com>

ref #827 #955

Normalize the dimension CRS (e.g. "EPSG:4326") before comparing it with the bbox CRS, so equivalent CRS notations no longer fall back to a non-concrete extent. Add `normalize_crs_or_passthrough` helper that leaves unparseable CRS values untouched instead of raising. ref #827 #955

to make user-facing occurrences more self-explanatory ref #827 #955

ref #827 #955

soxofaan force-pushed the issue827-cube-metadata-viz branch from c24bb72 to fbbf322 Compare October 9, 2026 13:01

soxofaan changed the title ~~Rudimentary _repr_html_ of cube dimension metadata~~ Pretty _repr_html_ of cube dimension metadata Oct 9, 2026

soxofaan requested a balanced review from Copilot October 9, 2026 21:47

Copilot started reviewing on behalf of soxofaan October 9, 2026 21:47

View session

Copilot AI reviewed Oct 9, 2026

View reviewed changes

soxofaan force-pushed the issue827-cube-metadata-viz branch from 54f4316 to 9b9f13f Compare October 12, 2026 10:19

soxofaan requested a balanced review from Copilot October 12, 2026 10:20

Copilot started reviewing on behalf of soxofaan October 12, 2026 10:20

View session

Copilot AI reviewed Oct 12, 2026

View reviewed changes

Comment thread openeo/utils/datetime.py Resolved

Comment thread openeo/metadata/__init__.py Resolved

soxofaan force-pushed the issue827-cube-metadata-viz branch 2 times, most recently from 281ce2b to a2ab17a Compare October 12, 2026 13:16

soxofaan requested a balanced review from Copilot October 12, 2026 13:16

Copilot started reviewing on behalf of soxofaan October 12, 2026 13:17

View session

Copilot AI reviewed Oct 12, 2026

View reviewed changes

Comment thread openeo/metadata/__init__.py Resolved

Comment thread openeo/metadata/__init__.py Resolved

Comment thread openeo/utils/datetime.py Resolved

Comment thread openeo/utils/datetime.py Resolved

soxofaan added a commit that referenced this pull request Oct 12, 2026

Handle filter_temporal end-exclusiveness in DateTimeInterval #955

811d9fa

soxofaan requested a balanced review from Copilot October 12, 2026 16:37

Copilot started reviewing on behalf of soxofaan October 12, 2026 16:38

View session

Copilot AI reviewed Oct 12, 2026

View reviewed changes

Comment thread openeo/metadata/__init__.py Resolved

soxofaan added a commit that referenced this pull request Oct 12, 2026

fixup! Handle filter_temporal end-exclusiveness in DateTimeInterval #…

7eff22a

…955

soxofaan added a commit that referenced this pull request Oct 12, 2026

fixup! Handle filter_temporal end-exclusiveness in DateTimeInterval #…

cc94f24

…955

soxofaan added a commit that referenced this pull request Oct 12, 2026

fixup! fixup! Handle filter_temporal end-exclusiveness in DateTimeI…

3934dab

…nterval #955

soxofaan added a commit that referenced this pull request Oct 12, 2026

Handle filter_temporal end-exclusiveness in DateTimeInterval #955

e288323

soxofaan force-pushed the issue827-cube-metadata-viz branch from 3934dab to e288323 Compare October 12, 2026 19:12

soxofaan requested a balanced review from Copilot October 12, 2026 19:17

Copilot started reviewing on behalf of soxofaan October 12, 2026 19:18

View session

Copilot AI reviewed Oct 12, 2026

View reviewed changes

Comment thread openeo/metadata/__init__.py Resolved

soxofaan added a commit that referenced this pull request Oct 13, 2026

fixup! Handle filter_temporal end-exclusiveness in DateTimeInterval #…

48dac2c

…955

soxofaan requested a balanced review from Copilot October 13, 2026 07:01

Copilot started reviewing on behalf of soxofaan October 13, 2026 07:02

View session

soxofaan added a commit that referenced this pull request Oct 23, 2026

Handle filter_temporal end-exclusiveness in DateTimeInterval #955

122be58

soxofaan added a commit that referenced this pull request Oct 23, 2026

Add axis attribute to SpatialDimension

f21697f

related to #955, #827

Copilot AI reviewed Oct 23, 2026

View reviewed changes

soxofaan commented Oct 28, 2026

View reviewed changes

Comment thread openeo/metadata/_extents.py Resolved

Copilot AI reviewed Oct 28, 2026

View reviewed changes

soxofaan commented Oct 28, 2026

View reviewed changes

Comment thread openeo/metadata/_jupyter.py Resolved

soxofaan added 5 commits October 29, 2026 11:10

Initial table-based repr_html of cube dimension metadata #827

2b4a59d

Handle filter_temporal end-exclusiveness in DateTimeInterval #955

8b42e08

Add axis attribute to SpatialDimension

41f48d6

related to #955, #827

Reorganise test_metadata.py

ccb7099

(re)group test functions in relevant test classes for better overview and easier future evolution refs #949 #827 #955

soxofaan commented Oct 29, 2026

View reviewed changes

Comment thread CHANGELOG.md Resolved

soxofaan and others added 10 commits October 29, 2026 13:45

Basic handling of filter_bbox in CubeMetadata, down to SpatialDimension

59f382c

refs #827, #949

Exclude antimeridian handling from metadata filter_bbox for now

d2bb21e

refs #827, #949

minor #867 fixup

67338ae

Track SpatialDimension.crs from _StacMetadataParser

143301b

ref #827 #955

Ignore "z" spatial dimension in CubeMetadata

3f5539f

ref #827 #955

Potential fix for pull request finding 'filter_bbox crashes on x/y/z …

5895ae3

…cubes with z dimension' Co-authored-by: Copilot Autofix powered by AI <175728472+Copilot@users.noreply.github.com>

Improve default crs handling in SpatialDimension

ad7be78

ref #827 #955

add a reason to _NonConcreteExtent

b23fe87

to make user-facing occurrences more self-explanatory ref #827 #955

cube metadata table in jupyter: force left text alignment

4fa3ae3

ref #827 #955

soxofaan mentioned this pull request Oct 29, 2026

unit test failures due to cube metadata changes Open-EO/openeo-geopyspark-driver#1927

Closed

EmileSonneveld mentioned this pull request Oct 29, 2026

Use has_spatial_dimensions to fix test_extract_result_metadata_aggreg… Open-EO/openeo-python-driver#554

Draft

Conversation

soxofaan commented Oct 9, 2026

Uh oh!

soxofaan commented Oct 9, 2026

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

🟡 Changes recommended

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

🟡 Changes recommended

Uh oh!

Uh oh!

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

🟡 Changes recommended

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

🟡 Changes recommended

Uh oh!

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

🟡 Changes recommended

Uh oh!

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

🔵 Needs a closer look

Uh oh!

soxofaan commented Oct 26, 2026

Uh oh!

soxofaan commented Oct 28, 2026

Uh oh!

Uh oh!

soxofaan commented Oct 28, 2026

Uh oh!

Copilot AI commented Oct 28, 2026

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

🟢 Approval recommended

Uh oh!

Uh oh!

soxofaan commented Oct 28, 2026

Uh oh!

soxofaan commented Oct 28, 2026

Uh oh!

Uh oh!

Reviewers

Uh oh!

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants