Interactive downsampling large vector fields (example gallery) #1372
Conversation
|
Hi @iuryt, it looks like pre-commit has caught that the notebook hasn't been cleaned up before being pushed to Github. You need to install pre-commit as described in the developer guide (and run it manually with |
|
Sorry. Gonna fix that. |
Codecov ReportAll modified and coverable lines are covered by tests ✅
Additional details and impacted files@@ Coverage Diff @@
## main #1372 +/- ##
==========================================
+ Coverage 87.39% 88.80% +1.40%
==========================================
Files 50 51 +1
Lines 7490 7510 +20
==========================================
+ Hits 6546 6669 +123
+ Misses 944 841 -103 ☔ View full report in Codecov by Sentry. |
|
@maximlt |
| "\n", | ||
| "# Create interactive plot components\n", | ||
| "range_xy = hv.streams.RangeXY() # Stream to capture range changes\n", | ||
| "filtered = hv.DynamicMap(create_quiver, streams=[range_xy]) # Dynamic plot\n", |
There was a problem hiding this comment.
Name it perhaps downsampled_quiver?
| "cell_type": "markdown", | ||
| "metadata": {}, | ||
| "source": [ | ||
| "If we just try to call `ds.hvplot.vectorfield` this is probably returning `MemoryError`. The alternative is to dynamically downsample the data based on the visible range. This helps manage memory consumption when dealing with large datasets, especially when plotting vector fields." |
There was a problem hiding this comment.
I would be nice to link to a HoloViews user guide and explain that we're going to use some HoloViews concepts to work around the memory error, while stile using hvPlot to create the quiver.
Co-authored-by: Maxime Liquet <[email protected]>
Merge from main
|
Thanks @maximlt ! |
I made a promise almost a year ago to create this example and here I am.
The idea came from those discussions from discourse.
https://discourse.holoviz.org/t/highres-quiver-plot-returning-memory-error/6261/26
https://discourse.holoviz.org/t/dealing-high-resolution-vector-plots-xarray-holoviz/6310
Direct plotting large vector fields can quickly consume excessive memory, leading to
MemoryError. To address this issue, we introduce a dynamic downsampling technique that enables interactive exploration of vector fields without sacrificing performance.