feat: add datafusion scalar UDF examples #9841
Draft
+2,558
−242
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
jleibs
reviewed
Apr 29, 2025
| *prior_columns, | ||
| bounded_image_extraction( | ||
| col(frame_column), | ||
| col(f"{boxes_2d_path}:Position2D"), |
There was a problem hiding this comment.
Very related to: #9837
I don't think we want to depend on hard-coded column names here.
Would be great if we could build a helper with a signature like:
resolve_component_selector(df.schema(), boxes_2d_path, rr.archetypes.Boxes2D.centers)
with some magic that uses rr.archetypes.Boxes2D.centers to resolve the full triple (archetype, component, field), and then scans the schema to find the column that's the best match
timsaucer
force-pushed
the
tsaucer/datafusion-scalar-udfs
branch
from
857bace to
4528429
Compare
|
Web viewer failed to build.
Note: This comment is updated whenever you push a commit. |
… a rrd that already exists instead of the manually extracted one.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This PR must remain in draft until the upstream datafusion version 48 is released
Overview
This PR adds in two DataFusion Scalar UDFs as the first building blocks of our analysis toolkit.
BoundedImageExtractionUdfwill take a video blob, frame ID, and 2d bounding boxes and extract the images within the bounding boxDepthImageToPointCloudUdfwill take a depth image and pinhole and extract a point cloudAlso provided are the wrapper functions that can be used to take a DataFusion DataFrame and pass user friendly entity paths to perform the required UDF manipulations.
There are two example jupyter notebooks demonstrating this using data from our standard examples.
Running the examples
First you will need to download the associated RRD for the demo files indicated in each notebook.
This requires a version of
datafusion-pythonthat is not yet released. You can pull from this branch and you will need to perform a local build usingAfter setting up the
uvenvironment per the readme in the repo, build:The in your
rerunrepo you will needYour exact .whl file will likely be different.
With that installed, you should be able to run the example jupyter notebooks.