Optimise when we get access to b-tree by providing lazier view of datasets, access to b-tree location, and new p5dump#138

Open

bnlawrence wants to merge 19 commits intomainfrom

Commits on Oct 21, 2025

Calculate b-tree range
Bryan Lawrence
committed
Merge remote-tracking branch 'origin/main' into btreeloc
Bryan Lawrence
committed

Commits on Oct 22, 2025

Remove redudant chunk handling from btree.py (part of #131)
Bryan Lawrence
committed
Just making a note that we need to be careful about chunk indexing should we have a v3 layout in the future.
Bryan Lawrence
committed
First cut at supporting ncdump like behaviour. Doesn't have support for file or group attributes yet, or phony dimensions.
Bryan Lawrence
committed

Commits on Oct 23, 2025

Working implementation of lazy access to variables (#135) and partial implementation of p5dump functionality (#134). Unit tests are failing due to a desire to get closer to (but not exactly) what n…
Bryan Lawrence
committed

Commits on Oct 24, 2025

allow visititems to be lazy
Bryan Lawrence
committed
p5dump works for the test cases
Bryan Lawrence
committed
Fixed string handling in groups
Bryan Lawrence
committed
Better testing
Bryan Lawrence
committed

Commits on Oct 26, 2025

Support for chunk information in p5dump via -s. (Fixed btree_range as well I think)
Bryan Lawrence
committed

Commits on Oct 27, 2025

p5dump -s includes storage type (from layout_class)
Bryan Lawrence
committed
Edge case detection and bug fix
Bryan Lawrence
committed
Handling broken pipes more gracefully
Bryan Lawrence
committed
Why don't I run my tests before committing?
Bryan Lawrence
committed
Cleaning up the DatasetID interface error handling for chunk queries on unchunked data and some tests to keep V happy.
Bryan Lawrence
committed
Merge remote-tracking branch 'origin/main' into optimise
Bryan Lawrence
committed
Merge branch 'main' into optimise
valeriupredoi
committed

Commits on Oct 28, 2025

Better checking of chunk info testing answers, courtesy of @zequihg50
Bryan Lawrence
committed