Adds a new brencher profile hermes_voice_gateway that deploys the voice-gateway compose from Org1/hermes-voice-gateway (mirrored in hermes/hermes-voice-gateway). Image is built by Gitea Actions on tag push and published to Gitea Packages.
Why no buildDocker
buildDocker=None because the voice-gateway image is built externally by .gitea/workflows/release.yml in Org1/hermes-voice-gateway. Brencher only needs to deploy. Image registry: gitea.rudolf.keenetic.link/hermes/hermes-voice-gateway.
Why version is hardcoded
Other stacks use auto-<commit_sha> because brencher builds images locally and tags them with the merge commit SHA. Voice-gateway image is published under a semver tag (v0.1.1) by Gitea Actions. To avoid publishing redundant auto-<sha> tags on every push, this stack hardcodes version = "v0.1.1".
To bump the version: update VOICE_GATEWAY_VERSION in hermes_voice_gateway.py AND the image tag in Org1/hermes-voice-gateway/compose.yaml.
Required upstream changes
Merge admin/ansible PR #4 (adds hermes-proxy service to hermes_infra stack).
Wire voice_gateway_models Docker volume with piper TTS voice model (ru_RU-denis-medium.onnx + .onnx.json).
Add hermes_voice_gateway to PROFILES env of brencher_brencher-backend service.
Confirm the stack name hermes_voice_gateway matches brencher's stack namespace handling.
Verification
docker stack ls shows stack hermes_voice_gateway.
docker service ps hermes_voice_gateway_voice-gateway shows task with image gitea.rudolf.keenetic.link/hermes/hermes-voice-gateway:v0.1.1.
## Summary
Adds a new brencher profile `hermes_voice_gateway` that deploys the voice-gateway compose from `Org1/hermes-voice-gateway` (mirrored in `hermes/hermes-voice-gateway`). Image is built by Gitea Actions on tag push and published to Gitea Packages.
## Why no buildDocker
`buildDocker=None` because the voice-gateway image is built externally by `.gitea/workflows/release.yml` in `Org1/hermes-voice-gateway`. Brencher only needs to deploy. Image registry: `gitea.rudolf.keenetic.link/hermes/hermes-voice-gateway`.
## Why version is hardcoded
Other stacks use `auto-<commit_sha>` because brencher builds images locally and tags them with the merge commit SHA. Voice-gateway image is published under a semver tag (`v0.1.1`) by Gitea Actions. To avoid publishing redundant `auto-<sha>` tags on every push, this stack hardcodes `version = "v0.1.1"`.
To bump the version: update `VOICE_GATEWAY_VERSION` in `hermes_voice_gateway.py` AND the image tag in `Org1/hermes-voice-gateway/compose.yaml`.
## Required upstream changes
1. Merge `admin/ansible` PR #4 (adds `hermes-proxy` service to `hermes_infra` stack).
2. Wire `voice_gateway_models` Docker volume with piper TTS voice model (`ru_RU-denis-medium.onnx` + `.onnx.json`).
3. Add `hermes_voice_gateway` to `PROFILES` env of `brencher_brencher-backend` service.
4. Confirm the stack name `hermes_voice_gateway` matches brencher's stack namespace handling.
## Verification
- `docker stack ls` shows stack `hermes_voice_gateway`.
- `docker service ps hermes_voice_gateway_voice-gateway` shows task with image `gitea.rudolf.keenetic.link/hermes/hermes-voice-gateway:v0.1.1`.
- `curl http://voice-gateway.keenetic.link:8765/healthz` returns `{"status": "ok"}`.
## Related
- Org1/hermes-voice-gateway#4 (voice-gateway service + browser client)
- hermes/ansible#4 (hermes-proxy service in hermes_infra)
- admin/ansible#7 (Gitea ROOT_URL fix, merged)
Deploys the voice-gateway compose from Org1/hermes-voice-gateway (fork
in hermes/hermes-voice-gateway). Image is built by Gitea Actions on tag
push and published to gitea.rudolf.keenetic.link/hermes/hermes-voice-gateway
under semver tags, so buildDocker=None and version is hardcoded to v0.1.1
instead of auto-<sha>. Bumping the release means updating this constant
plus the compose.yaml image tag in Org1/hermes-voice-gateway.
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
Summary
Adds a new brencher profile
hermes_voice_gatewaythat deploys the voice-gateway compose fromOrg1/hermes-voice-gateway(mirrored inhermes/hermes-voice-gateway). Image is built by Gitea Actions on tag push and published to Gitea Packages.Why no buildDocker
buildDocker=Nonebecause the voice-gateway image is built externally by.gitea/workflows/release.ymlinOrg1/hermes-voice-gateway. Brencher only needs to deploy. Image registry:gitea.rudolf.keenetic.link/hermes/hermes-voice-gateway.Why version is hardcoded
Other stacks use
auto-<commit_sha>because brencher builds images locally and tags them with the merge commit SHA. Voice-gateway image is published under a semver tag (v0.1.1) by Gitea Actions. To avoid publishing redundantauto-<sha>tags on every push, this stack hardcodesversion = "v0.1.1".To bump the version: update
VOICE_GATEWAY_VERSIONinhermes_voice_gateway.pyAND the image tag inOrg1/hermes-voice-gateway/compose.yaml.Required upstream changes
admin/ansiblePR #4 (addshermes-proxyservice tohermes_infrastack).voice_gateway_modelsDocker volume with piper TTS voice model (ru_RU-denis-medium.onnx+.onnx.json).hermes_voice_gatewaytoPROFILESenv ofbrencher_brencher-backendservice.hermes_voice_gatewaymatches brencher's stack namespace handling.Verification
docker stack lsshows stackhermes_voice_gateway.docker service ps hermes_voice_gateway_voice-gatewayshows task with imagegitea.rudolf.keenetic.link/hermes/hermes-voice-gateway:v0.1.1.curl http://voice-gateway.keenetic.link:8765/healthzreturns{"status": "ok"}.Related
Pull request closed