build: add stable-diffusion server to musa and vulkan container images

[mostlygeek]

Summary

• Adds sd-server from stable-diffusion.cpp container images to musa and vulkan variants
• llama-swap-sd.Containerfile is a thin 10-line layer on top of the base image (instead of duplicating the full 49-line Containerfile)
• Build script builds the base image first, then layers sd-server on top for supported architectures
• Includes build script hardening (set -euo pipefail, env var defaults, ${GITHUB_TOKEN:-} fix)
• Adds sd-server example to config.example.yaml

Supersedes #450 with a DRY approach.

Summary by CodeRabbit

• New Features

  • Added Stable Diffusion server integration for supported architectures.
  • Included a new example model configuration for Stable Diffusion usage.

• Chores

  • Improved build reliability with stricter error handling and safer defaults.
  • Added conditional build steps to produce an additional SD image layer for supported platforms.

_{✏️ Tip: You can customize this high-level summary in your review settings.}

[coderabbitai]

Warning

Rate limit exceeded

@mostlygeek has exceeded the limit for the number of commits that can be reviewed per hour. Please wait 18 minutes and 16 seconds before requesting another review.

⌛ How to resolve this issue?

After the wait time has elapsed, a review can be triggered using the @coderabbitai review command as a PR comment. Alternatively, push new commits to this PR.

We recommend that you space out your commits to avoid hitting the rate limit.

🚦 How do rate limits work?

CodeRabbit enforces hourly rate limits for each developer per organization.

Our paid plans have higher rate limits than the trial, open-source and free plans. In all cases, we re-allow further reviews after a brief timeout.

Please see our FAQ for further information.

Walkthrough

Adds Stable Diffusion server layering to the container build flow: strict bash error handling and debug flags in the build script, new SD_IMAGE/SD_TAG build args, a conditional post-build step that builds an sd-server layer for musa and vulkan architectures, a new Containerfile to copy the sd-server binary into the final image, and an example config entry for a "z-image" model.

Changes

Cohort / File(s)	Summary
Build script docker/build-container.sh	Adds set -euo pipefail, LOG_DEBUG/DEBUG_ABORT_BUILD and log_debug, defensive GITHUB_TOKEN handling, new SD_IMAGE/SD_TAG variables, and an architecture conditional that builds/pushes an additional sd-server image layer for musa/vulkan using llama-swap-sd.Containerfile.
Containerfile docker/llama-swap-sd.Containerfile	New multi-stage Containerfile that accepts SD_IMAGE/SD_TAG, pulls the SD image as a source stage, and copies /sd-server into the final llama-swap image with configurable UID/GID.
Example config docker/config.example.yaml	Adds a new models.z-image entry including checkEndpoint/cmd for /app/sd-server and an aliases list (formatting and new model declaration).

Estimated code review effort

🎯 3 (Moderate) | ⏱️ ~20 minutes

Possibly related PRs

• feat: improve llama.cpp base image tag for cpu #391: Overlaps with changes to docker/build-container.sh image/tag/build-arg logic.
• feat: Add support for custom llama.cpp base image and forked llama-swap repositories #396: Related changes to container build pipeline and build-arg handling for layered images.

🚥 Pre-merge checks | ✅ 2 | ❌ 1

❌ Failed checks (1 warning)

Check name	Status	Explanation	Resolution
Docstring Coverage	⚠️ Warning	Docstring coverage is 0.00% which is insufficient. The required threshold is 80.00%.	Write docstrings for the functions missing them to satisfy the coverage threshold.

✅ Passed checks (2 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title 'build: add stable-diffusion server to musa and vulkan container images' directly and specifically describes the main change: adding stable-diffusion server support to container images for two specific architectures. It is clear, concise, and accurately summarizes the primary objective of the changeset.

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

✨ Finishing touches

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Post copyable unit tests in a comment
• Commit unit tests in branch build/dry-sd-containerfile

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

[coderabbitai]

Actionable comments posted: 2

🤖 Fix all issues with AI agents

In `@docker/build-container.sh`:
- Line 114: The SD_TAG variable is set to an architecture-specific tag (SD_TAG)
that doesn't exist for musa/vulkan and causes the COPY --from stage to fail;
update the SD_TAG assignment (the SD_TAG variable used by the COPY --from step)
to use an upstream-compatible tag format (e.g.,
master-<build-number>-<git-short-sha>) or implement a lookup to fetch a valid
master-* tag from ghcr.io/leejet/stable-diffusion.cpp before the COPY --from
step, or alternatively ensure you build/publish architecture-specific images
yourself so the COPY --from stage can pull them; adjust any places referencing
SD_TAG accordingly.

In `@docker/config.example.yaml`:
- Around line 20-33: The z-image model entry is missing the proxy field and uses
a literal block for cmd which preserves newlines; add a proxy:
"http://127.0.0.1:9999" key to the z-image mapping and convert the cmd from a
literal block (|) to a folded block (>) so the command becomes a single-line
string (ensure all flags and paths for /app/sd-server, --listen-port,
--diffusion-model, --vae, --llm, --offload-to-cpu, --cfg-scale,
--height/--width, --steps remain space-separated) so llama-swap can route to the
spawned server and parse the command correctly.

[rare-magma]

@mostlygeek perhaps it would be worth adding some info to the readme to improve discoverability of this feature?

[mostlygeek]

@rare-magma oh yes, that's a good idea. Would you mind opening a PR with the config changes you introduced? Those were good.