Repository navigation
Patch Bookworm-compatible rover dependencies - #12
Conversation
|
Warning Amazon Q Developer for GitHub: end of support On January 30, 2027, AWS will discontinue support for the Amazon Q Developer for GitHub App. New app installations will no longer be available starting November 18, 2026. For capabilities similar to the Amazon Q Developer for GitHub App, explore Kiro and the Kiro GitHub App. For more information about the end of support, see Amazon Q Developer end-of-support announcement on the AWS DevOps Blog. |
There was a problem hiding this comment.
Review Summary
This PR updates requirements.txt to Bookworm-compatible versions for Python 3.11.2 on ARM64. Most of the version bumps are straightforward and appear correct. However, there are two blocking defects in the newly introduced PyTorch conditional dependency block that will prevent successful installation:
-
Invalid
===pip operator (lines 205, 207, 209) —===is not a recognised pip version specifier. Any install on a non-aarch64 platform will fail with anInvalidRequirementerror. These must be changed to==. -
torchaudioversion mismatch (line 206) —torchaudio==2.11.1is pinned againsttorch==2.13.0, but torchaudio must match the torch major+minor version. This will cause a runtimeRuntimeErroron aarch64 devices. It should betorchaudio==2.13.0+cpu.
Both issues need to be resolved before this PR can be merged.
You can now have the agent implement changes and create commits directly on your pull request's source branch. Simply comment with /q followed by your request in natural language to ask the agent to make changes.
| torch===2.13.0; sys_platform != "linux" or platform_machine != "aarch64" | ||
| torchaudio==2.11.1+cpu; sys_platform == "linux" and platform_machine == "aarch64" | ||
| torchaudio===2.11.1; sys_platform != "linux" or platform_machine != "aarch64" | ||
| torchvision==0.28.0+cpu; sys_platform == "linux" and platform_machine == "aarch64" | ||
| torchvision===0.28.0; sys_platform != "linux" or platform_machine != "aarch64" |
There was a problem hiding this comment.
🛑 Logic Error: The === operator is not valid pip syntax. pip only recognises ==, !=, <=, >=, <, >, and ~= as version specifiers. Using === will cause pip install to raise a InvalidRequirement error and abort the installation entirely for any platform that is not aarch64 Linux. All three non-aarch64 fallback lines must use == instead.
| torch===2.13.0; sys_platform != "linux" or platform_machine != "aarch64" | |
| torchaudio==2.11.1+cpu; sys_platform == "linux" and platform_machine == "aarch64" | |
| torchaudio===2.11.1; sys_platform != "linux" or platform_machine != "aarch64" | |
| torchvision==0.28.0+cpu; sys_platform == "linux" and platform_machine == "aarch64" | |
| torchvision===0.28.0; sys_platform != "linux" or platform_machine != "aarch64" | |
| torch==2.13.0; sys_platform != "linux" or platform_machine != "aarch64" | |
| torchaudio==2.11.1; sys_platform != "linux" or platform_machine != "aarch64" | |
| torchvision==0.28.0; sys_platform != "linux" or platform_machine != "aarch64" |
| --find-links https://download.pytorch.org/whl/cpu/torchvision/ | ||
| torch==2.13.0+cpu; sys_platform == "linux" and platform_machine == "aarch64" | ||
| torch===2.13.0; sys_platform != "linux" or platform_machine != "aarch64" | ||
| torchaudio==2.11.1+cpu; sys_platform == "linux" and platform_machine == "aarch64" |
There was a problem hiding this comment.
🛑 Logic Error: torchaudio version 2.11.1 is not compatible with torch==2.13.0. The PyTorch project requires torchaudio and torch to share the same major+minor version. Pairing torch 2.13.0 with torchaudio 2.11.1 will produce import errors at runtime (RuntimeError: torchaudio requires PyTorch 2.11.1). The correct version is torchaudio==2.13.0.
| torchaudio==2.11.1+cpu; sys_platform == "linux" and platform_machine == "aarch64" | |
| torchaudio==2.13.0+cpu; sys_platform == "linux" and platform_machine == "aarch64" |
Update the rover's web, notebook, parsing, TLS, and vision/audio dependencies to patched published releases that resolve on Raspberry Pi OS Bookworm with Python 3.11.2. The ARM64 PyTorch packages use the upstream CPU wheels. The selected MediaPipe release retains the existing hand, face, pose and drawing APIs.
Both the original and candidate environments passed all 181 rover tests on isolated Bookworm Python 3.11.2 without native rover devices or external network. Both passed camera API, CPU Torch, Flask/Socket.IO, notebook and crypto smoke tests. The candidate resolved 146 packages for ARM64. The optional Bookworm system Matplotlib preset reproduced the same camera import error on both environments. No native Pi or GPU run was performed.
Four advisories remain within the supported platform constraints. One Protobuf fix needs 5.29.6, beyond the legacy MediaPipe range, while the Debian backport exposes Python version 4.21.12, below the minimum of the newer compatible MediaPipe releases. Fixed Soup Sieve wheels require at least Python 3.11.5.