Describe the bug
Attempting to log a partition'd TensorDataBatch results in an error.
To Reproduce
Steps to reproduce the behavior:
# file: rerun_issue.py
import numpy as np
import rerun as rr
batch, dim = 10, 20
array = np.zeros((batch, dim))
indices = np.arange(batch)
lengths = np.array([dim] * batch)
rr.send_columns(
"/entity",
[rr.TimeSequenceColumn("indices", indices)],
[rr.components.TensorDataBatch(array).partition(lengths)],
strict=True,
)Expected behavior
Logging succeeds.
Backtrace
uv run --isolated --with rerun-sdk==0.20.3 rerun_issue.py
Traceback (most recent call last):
File "/Users/<user>/rerun_issue.py", line 11, in <module>
rr.send_columns(
File "/Users/<user>/Library/Caches/uv/archive-v0/cvPpsKaWxYZeojMA2lgt_/lib/python3.12/site-packages/rerun_sdk/rerun/error_utils.py", line 191, in wrapper
return func(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^
File "/Users/<user>/Library/Caches/uv/archive-v0/cvPpsKaWxYZeojMA2lgt_/lib/python3.12/site-packages/rerun_sdk/rerun/_send_columns.py", line 234, in send_columns
arrow_list_array = component_column.as_arrow_array()
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/Users/<user>/Library/Caches/uv/archive-v0/cvPpsKaWxYZeojMA2lgt_/lib/python3.12/site-packages/rerun_sdk/rerun/_baseclasses.py", line 310, in as_arrow_array
return pa.ListArray.from_arrays(offsets, array)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "pyarrow/array.pxi", line 2557, in pyarrow.lib.ListArray.from_arrays
File "pyarrow/array.pxi", line 1686, in pyarrow.lib.Array.validate
File "pyarrow/error.pxi", line 92, in pyarrow.lib.check_status
pyarrow.lib.ArrowInvalid: Length spanned by list offsets (200) larger than values array (length 1)
Desktop (please complete the following information):
Rerun version
rerun-cli 0.20.3 (default map_view nasm native_viewer release web_viewer) [rustc 1.79.0 (129f3b996 2024-06-10), LLVM 18.1.7] aarch64-apple-darwin release-0.20.3 7a031f3,
built 2024-12-03T17:44:19Z
Video features: av1 default ffmpeg nasm serde
Additional context
This issue doesn't occur for at least some other batch types (e.g. ScalarBatch), or without .partition(...).
Describe the bug
Attempting to log a
partition'dTensorDataBatchresults in an error.To Reproduce
Steps to reproduce the behavior:
Expected behavior
Logging succeeds.
Backtrace
Desktop (please complete the following information):
Rerun version
rerun-cli 0.20.3 (default map_view nasm native_viewer release web_viewer) [rustc 1.79.0 (129f3b996 2024-06-10), LLVM 18.1.7] aarch64-apple-darwin release-0.20.3 7a031f3,
built 2024-12-03T17:44:19Z
Video features: av1 default ffmpeg nasm serde
Additional context
This issue doesn't occur for at least some other batch types (e.g.
ScalarBatch), or without.partition(...).