Skip to content

Add pipelines to python interface - #223

Merged
phapalova merged 13 commits into
mainfrom
phapalova/python-pipelines
Aug 5, 2026
Merged

Add pipelines to python interface#223
phapalova merged 13 commits into
mainfrom
phapalova/python-pipelines

Conversation

@phapalova

Copy link
Copy Markdown
Collaborator

No description provided.

Signed-off-by: Petra Hapalova <phapalova@nvidia.com>
Parse custom scene params JSON without exceptions so malformed input is
rejected instead of aborting the process on Windows, and guard the C entry
point against anything escaping across the ABI boundary.

Size the default mesh color test off the positions helper rather than running
a native VoxelBVH build on a device-less session, which never completed on
CPU-only runners, and shrink the RGBA8 bake inputs so software-Vulkan runners
do not run out of memory.

Signed-off-by: Petra Hapalova <phapalova@nvidia.com>
Reduce Gaussian point count, raster3d voxel size, and VoxelBVH resolution so
conversion tests stay within runner memory on lavapipe/MoltenVK.

Signed-off-by: Petra Hapalova <phapalova@nvidia.com>
@phapalova
phapalova force-pushed the phapalova/python-pipelines branch from 0657baf to 03210d7 Compare August 4, 2026 07:03
Merge the duplicate raster3d builds into one PipelineType path, drop
Gaussian point count and resolutions again, and bake RGBA8 from a tiny
mesh so software-Vulkan runners no longer OOM or crash mid-suite.

Signed-off-by: Petra Hapalova <phapalova@nvidia.com>
Gaussian raster/VoxelBVH and multi-direction RGBA8 bake keep OOMing or
hanging on software-Vulkan CI. Leave the light mesh conversions enabled
and allow forcing the heavy suite with NANOVDB_EDITOR_RUN_HEAVY_PIPELINE_TESTS=1.

Signed-off-by: Petra Hapalova <phapalova@nvidia.com>
Even tiny mesh VoxelBVH builds hang, OOM, or crash software-Vulkan CI
runners. Keep the registry and validation coverage, and allow forcing the
full conversion suite with NANOVDB_EDITOR_RUN_PIPELINE_CONVERSION_TESTS=1.

Signed-off-by: Petra Hapalova <phapalova@nvidia.com>
Signed-off-by: Petra Hapalova <phapalova@nvidia.com>
VoxelBVH mesh builds exceed the 300s lavapipe timeout there, matching the
existing streaming-smoke opt-out. Also keep the GitHub Actions smoke to a
single mesh build so Linux x64 does not sit in multi-build conversions.

Signed-off-by: Petra Hapalova <phapalova@nvidia.com>
Even the one-mesh smoke hangs on Linux lavapipe and access-violates on
Windows ARM64 inside nanovdb_from_triangles_array. Keep the light smoke on
macOS MoltenVK and leave the registry/validation coverage everywhere else.

Signed-off-by: Petra Hapalova <phapalova@nvidia.com>
Signed-off-by: Petra Hapalova <phapalova@nvidia.com>
Signed-off-by: Petra Hapalova <phapalova@nvidia.com>
Signed-off-by: Petra Hapalova <phapalova@nvidia.com>
@phapalova
phapalova marked this pull request as ready for review August 5, 2026 10:49
@phapalova
phapalova requested a review from areidmeyer as a code owner August 5, 2026 10:49
Signed-off-by: Petra Hapalova <phapalova@nvidia.com>
@phapalova
phapalova enabled auto-merge (squash) August 5, 2026 11:30
@phapalova
phapalova merged commit 120b2ed into main Aug 5, 2026
37 checks passed
@phapalova
phapalova deleted the phapalova/python-pipelines branch August 5, 2026 13:15
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants