nomyo-ai/nomyo-router
2
2
Fork 0
Code Issues 1 Pull requests Projects Releases 5 Packages 1 Wiki Activity Actions

Merge pull request 'dev-v0.7.x' (#21) from dev-v0.7.x into main
All checks were successful
Build and Publish Docker Image (Semantic Cache) / build (amd64, linux/amd64, docker-amd64) (push) Successful in 35s
Details
Build and Publish Docker Image / build (amd64, linux/amd64, docker-amd64) (push) Successful in 36s
Details
Build and Publish Docker Image (Semantic Cache) / build (arm64, linux/arm64, docker-arm64) (push) Successful in 10m21s
Details
Build and Publish Docker Image (Semantic Cache) / merge (push) Successful in 32s
Details
Build and Publish Docker Image / build (arm64, linux/arm64, docker-arm64) (push) Successful in 10m3s
Details
Build and Publish Docker Image / merge (push) Successful in 32s
Details

Browse source 
Reviewed-on: https://bitfreedom.net/code/code/nomyo-ai/nomyo-router/pulls/21
This commit is contained in:
Alpha Nerd 2026-04-08 14:01:54 +02:00
commit 07b80e654f
3 changed files with 11 additions and 8 deletions
Download patch file Download diff file Expand all files Collapse all files

11
.forgejo/workflows/docker-publish-semantic.yml
View file

@ -18,7 +18,6 @@ on:
env: env:
REGISTRY: bitfreedom.net REGISTRY: bitfreedom.net
IMAGE_NAME: ${{ github.repository }} IMAGE_NAME: ${{ github.repository }}
CACHE_IMAGE: ${{ github.repository }}-buildcache-semantic
DOCKER_BUILD_SUMMARY: "false" DOCKER_BUILD_SUMMARY: "false"
jobs: jobs:
@ -87,9 +86,9 @@ jobs:
provenance: false provenance: false
build-args: | build-args: |
SEMANTIC_CACHE=true SEMANTIC_CACHE=true
tags: ${{ env.REGISTRY }}/${{ env.CACHE_IMAGE }}:platform-semantic-${{ matrix.arch }} tags: ${{ env.REGISTRY }}/${{ env.IMAGE_NAME }}:semantic-platform-${{ matrix.arch }}
cache-from: type=registry,ref=${{ env.REGISTRY }}/${{ env.CACHE_IMAGE }}:buildcache-semantic-${{ matrix.arch }} cache-from: type=registry,ref=${{ env.REGISTRY }}/${{ env.IMAGE_NAME }}:buildcache-semantic-${{ matrix.arch }}
cache-to: type=registry,ref=${{ env.REGISTRY }}/${{ env.CACHE_IMAGE }}:buildcache-semantic-${{ matrix.arch }},mode=min cache-to: type=registry,ref=${{ env.REGISTRY }}/${{ env.IMAGE_NAME }}:buildcache-semantic-${{ matrix.arch }},mode=max
merge: merge:
runs-on: docker-amd64 runs-on: docker-amd64
@ -145,6 +144,6 @@ jobs:
run: | run: |
docker buildx imagetools create \ docker buildx imagetools create \
$(jq -cr '.tags | map("-t " + .) | join(" ")' <<< "$DOCKER_METADATA_OUTPUT_JSON") \ $(jq -cr '.tags | map("-t " + .) | join(" ")' <<< "$DOCKER_METADATA_OUTPUT_JSON") \
${{ env.REGISTRY }}/${{ env.CACHE_IMAGE }}:platform-semantic-amd64 \ ${{ env.REGISTRY }}/${{ env.IMAGE_NAME }}:semantic-platform-amd64 \
${{ env.REGISTRY }}/${{ env.CACHE_IMAGE }}:platform-semantic-arm64 ${{ env.REGISTRY }}/${{ env.IMAGE_NAME }}:semantic-platform-arm64

2
.forgejo/workflows/docker-publish.yml
View file

@ -79,7 +79,7 @@ jobs:
provenance: false provenance: false
tags: ${{ env.REGISTRY }}/${{ env.IMAGE_NAME }}:platform-${{ matrix.arch }} tags: ${{ env.REGISTRY }}/${{ env.IMAGE_NAME }}:platform-${{ matrix.arch }}
cache-from: type=registry,ref=${{ env.REGISTRY }}/${{ env.IMAGE_NAME }}:buildcache-${{ matrix.arch }} cache-from: type=registry,ref=${{ env.REGISTRY }}/${{ env.IMAGE_NAME }}:buildcache-${{ matrix.arch }}
cache-to: type=registry,ref=${{ env.REGISTRY }}/${{ env.IMAGE_NAME }}:buildcache-${{ matrix.arch }},mode=min cache-to: type=registry,ref=${{ env.REGISTRY }}/${{ env.IMAGE_NAME }}:buildcache-${{ matrix.arch }},mode=max
merge: merge:
runs-on: docker-amd64 runs-on: docker-amd64

6
router.py
View file

@ -2110,7 +2110,11 @@ async def chat_proxy(request: Request):
# Only cache when no max_tokens limit was set — otherwise # Only cache when no max_tokens limit was set — otherwise
# finish_reason=length might just mean max_tokens was hit, # finish_reason=length might just mean max_tokens was hit,
# not that the context window was exhausted. # not that the context window was exhausted.
_req_max_tok = params.get("max_tokens") or params.get("max_completion_tokens") or params.get("num_predict") _req_max_tok = (
params.get("max_tokens") or params.get("max_completion_tokens") or params.get("num_predict")
if use_openai else
(options.get("num_predict") if options else None)
)
if _dr == "length" and not _req_max_tok: if _dr == "length" and not _req_max_tok:
_pt = getattr(chunk, "prompt_eval_count", 0) or 0 _pt = getattr(chunk, "prompt_eval_count", 0) or 0
_ct = getattr(chunk, "eval_count", 0) or 0 _ct = getattr(chunk, "eval_count", 0) or 0