refactor(autoware_tensorrt_plugins): remove SegmentCSR nD indptr path

[mojomex]

Summary

Third PR in the SegmentCSR split stack.

This PR removes the unreachable n-dimensional indptr launcher path. The TensorRT plugin already requires a 2D source tensor and 1D indptr, so the vector-shaped launcher, indptr_to_offset helper and bookkeeping are dead code.

Since the SegmentCSR plugin has been introduces for PTv3, which isn't using the nD case, and since there are no other users, removing seems like the best option.

Stack

1. test(autoware_tensorrt_plugins): add SegmentCSR kernel tests #12740: add SegmentCSR kernel tests.
2. refactor(autoware_tensorrt_plugins): clarify SegmentCSR contract #12741: document/clarify the contract and clean up in/out API names.
3. This PR: remove the dead n-dimensional indptr path.
4. perf(autoware_tensorrt_plugins): keep SegmentCSR allocation-free #12555: keep SegmentCSR allocation-free.

[github-actions]

Thank you for contributing to the Autoware project!

🚧 If your pull request is in progress, switch it to draft mode.

Please ensure:

• You've checked our contribution guidelines.
• Your PR follows our pull request guidelines.
• All required CI checks pass before marking the PR ready for review.

[Copilot]

Pull request overview

This PR refactors the SegmentCSR TensorRT plugin and its CUDA kernel launcher to remove the previously-unreachable n-dimensional indptr/size-vector code path, aligning the implementation with the plugin’s documented 2D src + 1D indptr contract.

Changes:

• Simplifies segment_csr_launch API from std::vector<int32_t> size parameters to explicit (num_rows, num_cols, indptr_size) integers.
• Removes nD indptr bookkeeping utilities (including indptr_to_offset) and related kernel indexing paths.
• Updates plugin enqueue code and unit tests to use the new launcher signature.

Reviewed changes

Copilot reviewed 5 out of 5 changed files in this pull request and generated 3 comments.

Show a summary per file

File	Description
perception/autoware_tensorrt_plugins/test/segment_csr_kernel_test.cpp	Updates kernel test calls to the new segment_csr_launch signature.
perception/autoware_tensorrt_plugins/src/segment_csr_plugin.cpp	Updates plugin enqueue logic to pass (num_rows, num_cols, indptr_size) directly.
perception/autoware_tensorrt_plugins/src/scatter_ops/segment_csr.cu	Removes nD launch path, simplifies kernels/launch signature, and deletes nD offset logic.
perception/autoware_tensorrt_plugins/include/autoware/scatter_ops/utils.cuh	Removes the now-dead indptr_to_offset helper.
perception/autoware_tensorrt_plugins/include/autoware/scatter_ops/segment_csr.h	Replaces vector-based API with scalar size args and adds/updates contract documentation.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 0.00%. Comparing base (54fa737) to head (0f14e0a).
⚠️ Report is 4 commits behind head on main.

Additional details and impacted files

@@             Coverage Diff             @@
##             main   #12739       +/-   ##
===========================================
- Coverage   19.58%    0.00%   -19.59%     
===========================================
  Files        1903       41     -1862     
  Lines      131340     1587   -129753     
  Branches    45785        0    -45785     
===========================================
- Hits        25718        0    -25718     
+ Misses      84661     1587    -83074     
+ Partials    20961        0    -20961     

Flag	Coverage Δ
full-suite	0.00% <ø> (-19.59%)	⬇️

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[amadeuszsz]

@codex review

[chatgpt-codex-connector]

Codex Review: Didn't find any major issues. Bravo.

Reviewed commit: 0f14e0a9b7

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[amadeuszsz]

LGTM!