konflux: prefetch python dependencies when building -rag images

[mikebonnet]

Konflux supports prefetching dependencies via Hermeto. This allows for creation of more accurate SBOMs and source containers, and will improve the performance and reliability of builds when caching is implemented.

Retrieve dependencies for the -rag images using prefetch, and run the uv pip install command offline.

Summary by Sourcery

Enable prefetching Python dependencies for RAG container images using Hermeto and uv to generate accurate SBOMs and support offline pip installs in CI builds.

New Features:

• Generate and split requirements files per backend and index using uv pip compile and an AWK script to enable Hermeto prefetch.
• Add uv version check and incorporate offline pip install logic in build_rag.sh, including a fixup for gguf package archives.
• Inject prefetch-input configuration blocks into Tekton pipelines for CUDA, ROCm, Intel GPU, and CPU RAG images.

Enhancements:

• Refactor Makefiles to handle multiple Python versions and platforms, rename and combine requirement targets, and orchestrate generation of prefetched requirement files.
• Update root Makefile rag-requirements target to trigger all new prefetched requirement files.

[sourcery-ai]

Reviewer's Guide

This PR integrates Hermeto-based prefetching into the -rag image build process by overhauling Makefile rules for uv pip compilation, splitting and naming requirement files per index, updating the build script for offline installation with gguf fixups, and injecting prefetch configuration into Tekton pipelines.

File-Level Changes

Change	Details	Files
Overhaul Makefiles to drive uv-based prefetch and dynamic requirement compilation	Rename version/platform variables and derive per-backend values Introduce check-uv target enforcing minimum uv version Implement combined and split requirement targets with uv pip compile Update rag-requirements target and root Makefile invoke Parameterize Python version and platform per backend and architecture	container-images/common/Makefile Makefile
Add script to split compiled requirements by index source	Create split-reqs.awk to parse --emit-index-annotation output Generate per-index requirement files preserving headers Prevent manual edits by embedding generation comment	container-images/common/split-reqs.awk
Enhance build_rag.sh for offline installation and gguf repackaging	Assemble uv pip install arguments for multiple index files Comment out index-url lines and use PIP_FIND_LINKS with --offline Add fixup_gguf to repackage gguf tarball and update checksum Replace legacy uv pip sync calls and clean up temporary artifacts	container-images/scripts/build_rag.sh
Rename and introduce per-backend/index requirement files	Strip requirements-rag prefix and rename to backend.txt Add pypi.org and download.pytorch.org variants per backend and arch Create build.txt and build-cuda.txt for build-time dependencies Add requirements-rag-build.in input for build targets	container-images/common/requirements-rag.in container-images/common/requirements-rag-build.in container-images/common/build.txt container-images/common/build-cuda.txt container-images/common/cpu-pypi.org.txt container-images/common/cpu-download.pytorch.org.txt container-images/common/cpu-aarch64-pypi.org.txt container-images/common/cpu-aarch64-download.pytorch.org.txt container-images/common/cu128-pypi.org.txt container-images/common/cu128-download.pytorch.org.txt container-images/common/cu128-aarch64-pypi.org.txt container-images/common/cu128-aarch64-download.pytorch.org.txt container-images/common/rocm6.3-pypi.org.txt container-images/common/rocm6.3-download.pytorch.org.txt container-images/common/xpu-pypi.org.txt container-images/common/xpu-download.pytorch.org.txt
Inject Hermeto prefetch-input into Tekton pipeline specs	Add prefetch-input section with pip type and file lists Reference both requirements_files and requirements_build_files Enable binary caching directive for :all: packages	.tekton/cuda-rag/cuda-rag-pull-request.yaml .tekton/cuda-rag/cuda-rag-push.yaml .tekton/ramalama-rag/ramalama-rag-pull-request.yaml .tekton/ramalama-rag/ramalama-rag-push.yaml .tekton/intel-gpu-rag/intel-gpu-rag-pull-request.yaml .tekton/intel-gpu-rag/intel-gpu-rag-push.yaml .tekton/rocm-rag/rocm-rag-pull-request.yaml

---

Tips and commands

Interacting with Sourcery

• Trigger a new review: Comment @sourcery-ai review on the pull request.
• Continue discussions: Reply directly to Sourcery's review comments.
• Generate a GitHub issue from a review comment: Ask Sourcery to create an
  issue from a review comment by replying to it. You can also reply to a
  review comment with @sourcery-ai issue to create an issue from it.
• Generate a pull request title: Write @sourcery-ai anywhere in the pull
  request title to generate a title at any time. You can also comment
  @sourcery-ai title on the pull request to (re-)generate the title at any time.
• Generate a pull request summary: Write @sourcery-ai summary anywhere in
  the pull request body to generate a PR summary at any time exactly where you
  want it. You can also comment @sourcery-ai summary on the pull request to
  (re-)generate the summary at any time.
• Generate reviewer's guide: Comment @sourcery-ai guide on the pull
  request to (re-)generate the reviewer's guide at any time.
• Resolve all Sourcery comments: Comment @sourcery-ai resolve on the
  pull request to resolve all Sourcery comments. Useful if you've already
  addressed all the comments and don't want to see them anymore.
• Dismiss all Sourcery reviews: Comment @sourcery-ai dismiss on the pull
  request to dismiss all existing Sourcery reviews. Especially useful if you
  want to start fresh with a new review - don't forget to comment
  @sourcery-ai review to trigger a new review!

Customizing Your Experience

Access your dashboard to:

• Enable or disable review features such as the Sourcery-generated pull request
  summary, the reviewer's guide, and others.
• Change the review language.
• Add, remove or edit custom review instructions.
• Adjust other review settings.

Getting Help

• Contact our support team for questions or feedback.
• Visit our documentation for detailed guides and information.
• Keep in touch with the Sourcery team by following us on X/Twitter, LinkedIn or GitHub.