refactor(lidar_apollo_segmentation_tvm/_nodes): remove autoware_auto_common dependency

[xmfcx]

Description

Continuation of: #2234
Related to: #2229

Moved here to collaborate easier on AWF organization branch.

In order to build these packages, you need to pass --cmake-args -DDOWNLOAD_ARTIFACTS=1 to them.

I use colcon build --symlink-install --cmake-args -DCMAKE_BUILD_TYPE=RelWithDebInfo -DCMAKE_EXPORT_COMPILE_COMMANDS=1 -DDOWNLOAD_ARTIFACTS=1 --packages-select lidar_apollo_segmentation_tvm lidar_apollo_segmentation_tvm_nodes tvm_utility for convenience.

Summary

🤖 Generated by Copilot at 1f601f7

This pull request refactors the code and types of the lidar_apollo_segmentation_tvm package and its node, to use standard types instead of custom ones for the inference input and output data. This simplifies the code, improves the performance, and avoids unnecessary conversions and dependencies. It affects the files pipeline.hpp, lidar_apollo_segmentation_tvm_node.cpp, and all the files in the lidar_apollo_segmentation_tvm module and its test. It also updates the package.xml files to remove unused dependencies.

Walkthrough

🤖 Generated by Copilot at 1f601f7

• Replace bool8_t, char8_t, float32_t and float64_t types with standard types bool, int8_t, float and double in all files (F0-F13)
• Remove unnecessary dependencies and header files for autoware_auto_msgs and autoware_auto_common packages, which defined the replaced types (link, link, link, link, link, link, link, link, link, link, link, link, link, link, link)
• Add header file <cstdint> to include the definition of int8_t type, which is used for the model version (link, link, link)
• Modify the type of the version_from parameter in the tvm_utility::Pipeline constructor from std::array<char, 3> to std::array<int8_t, 3>, to match the type of the modelzoo_version field in the config_ object (link)
• Modify the types of the parameters and fields related to the point cloud data, such as pc_ptr, pos_x, pos_y, affine_matrix, offset, point_step, etc., from float32_t to float, to avoid unnecessary conversions and use a standard type (link, link, link, link, link)
• Modify the type of the epsilon variable in the FeatureGenerator::generate function from float64_t to double, to avoid unnecessary conversions and use a standard type (link)
• Modify the types of the fields related to the clustering algorithm, such as node_rank, traversed, is_center and is_object, from char8_t and bool8_t to int8_t and bool, to use standard types and avoid unnecessary conversions (link)
• Modify the type of the v vector in the main test function from uchar8_t to unsigned char, to use a standard type (link)

Poem

🤖 Generated by Copilot at 1f601f7

We're heaving away the old types, me hearties
We're making the code more simple and smart
We're using the standard float and bool, me hearties
We're matching the types with the inference part

Pre-review checklist for the PR author

The PR author must check the checkboxes below when creating the PR.

• I've confirmed the contribution guidelines.
• The PR follows the pull request guidelines.

In-review checklist for the PR reviewers

The PR reviewers must check the checkboxes below before approval.

• The PR follows the pull request guidelines.

Post-review checklist for the PR author

The PR author must check the checkboxes below before merging.

• There are no open discussions or they are tracked via tickets.

After all checkboxes are checked, anyone who has write access can merge the PR.

[xmfcx]

@ambroise-arm this way it compiles on my machine. Also used int8_t instead of char to keep things consistent.

Could you review it please?

[ambroise-arm]

I just want to double check if the char -> int8_t needs to be applied in other places as well (other Autoware packages and maybe modelzoo). I should be able to review later this week.

[xmfcx]

@ambroise-arm I've checked, colcon build --symlink-install --cmake-args -DCMAKE_BUILD_TYPE=RelWithDebInfo -DCMAKE_EXPORT_COMPILE_COMMANDS=1 -DDOWNLOAD_ARTIFACTS=1 --packages-select lidar_apollo_segmentation_tvm lidar_apollo_segmentation_tvm_nodes tvm_utility lidar_centerpoint_tvm compiles ok.

Which makes all the packages in the universe:

• lidar_apollo_segmentation_tvm
• lidar_apollo_segmentation_tvm_nodes
• lidar_centerpoint_tvm
• tvm_utility

I didn't check if they work in runtime though.

[ambroise-arm]

Generally looks good to me. I ran colcon test on the packages and it passes, so the inference doesn't fail catastrophically.
Just nitpicking, can you change char to int8_t in common/tvm_utility/include/tvm_utility/pipeline.hpp lines 204 (modelzoo_version) and 348 (version_up_to) as well please? For consistency.

[codecov]

Codecov Report

Patch and project coverage have no change.

Comparison is base (80385ab) 13.25% compared to head (cb1d451) 13.25%.

❗ Current head cb1d451 differs from pull request most recent head 21b629b. Consider uploading reports for the commit 21b629b to get more accurate results

Additional details and impacted files

@@           Coverage Diff           @@
##             main    #3136   +/-   ##
=======================================
  Coverage   13.25%   13.25%           
=======================================
  Files        1384     1384           
  Lines       97018    97017    -1     
  Branches    28213    28213           
=======================================
  Hits        12855    12855           
+ Misses      70354    70353    -1     
  Partials    13809    13809           

Flag	Coverage Δ		*Carryforward flag
differential	∅ <ø> (?)
total	13.25% <ø> (+<0.01%)	⬆️	Carriedforward from db7b5dd

*This pull request uses carry forward flags. Click here to find out more.

see 2 files with indirect coverage changes

Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here.

☔ View full report in Codecov by Sentry.
📢 Do you have feedback about the report comment? Let us know in this issue.

[xmfcx]

@ambroise-arm I've squashed+rebased the PR and applied the suggestions in cb1d451

Could you review again? (I've compiled it just to make sure)

[xmfcx]

Thank you for reviewing!