Repository navigation
Fix HDF5 variable attributes displayed as byte reprs - #372
Conversation
Codecov Report✅ All modified and coverable lines are covered by tests. Additional details and impacted files@@ Coverage Diff @@
## develop #372 +/- ##
===========================================
+ Coverage 97.08% 97.11% +0.02%
===========================================
Files 9 9
Lines 584 589 +5
===========================================
+ Hits 567 572 +5
Misses 17 17
Flags with carried forward coverage won't be shown. Click here to find out more. ☔ View full report in Codecov by Harness. 🚀 New features to boost your workflow:
|
|
@danielfromearth This HDF5 attribute decoding fix is now refreshed onto current develop, 0 behind, with all Python 3.11-3.14 tests and Codecov checks green and no unresolved review threads. It has been open since September 10 without a review. Could you review the current head when convenient? |
|
Thanks for another useful contribution, @sylvesterkaczmarek! This fixes the b'NASA' problem nicely. Good edge-case coverage too with non-ASCII, empty strings, the literal One small thing: |
|
Done in 6bf2c9f. I renamed the helper to value_to_comparable_str and updated the imports, internal calls, and test docstring so its cross-module use is explicit. git diff --check and Python compilation are clean; fresh hosted checks can validate the branch. |
Fixes #370.
Decode HDF5 scalar attribute bytes through the shared normalizer and decode string arrays before reference handling. Fixed- and variable-length storage now compare as the same text, including multidimensional arrays, while numeric arrays and object-reference resolution retain their existing behavior.
Validation on macOS arm64, Python 3.12.11:
git diff --checkpassed. Changelog updated.Live Earthdata downloads and Windows/Linux runs were not performed. The separate list-truncation issue #371 is outside this change.
📚 Documentation preview 📚: https://ncompare--372.org.readthedocs.build/en/372/