Use model cache for docker compose test (#1582)

Signed-off-by: ZePan110 <ze.pan@intel.com>
2025-03-04 09:48:27 +08:00
parent ce38a84372
commit e4de76da78
28 changed files with 61 additions and 31 deletions
--- a/.github/workflows/_example-workflow.yml
+++ b/.github/workflows/_example-workflow.yml
@@ -43,7 +43,11 @@ on:
      inject_commit:
        default: false
        required: false
-        type: string
+        type: boolean
+      use_model_cache:
+        default: false
+        required: false
+        type: boolean

 jobs:
 ####################################################################################################
@@ -110,6 +114,7 @@ jobs:
      tag: ${{ inputs.tag }}
      example: ${{ inputs.example }}
      hardware: ${{ inputs.node }}
+      use_model_cache: ${{ inputs.use_model_cache }}
    secrets: inherit


@@ -131,7 +136,7 @@ jobs:
 ####################################################################################################
  test-gmc-pipeline:
    needs: [build-images]
-    if: ${{ fromJSON(inputs.test_gmc) }}
+    if: false # ${{ fromJSON(inputs.test_gmc) }}
    uses: ./.github/workflows/_gmc-e2e.yml
    with:
      example: ${{ inputs.example }}
--- a/.github/workflows/_run-docker-compose.yml
+++ b/.github/workflows/_run-docker-compose.yml
@@ -28,6 +28,10 @@ on:
        required: false
        type: string
        default: ""
+      use_model_cache:
+        required: false
+        type: boolean
+        default: false
 jobs:
  get-test-case:
    runs-on: ubuntu-latest
@@ -144,9 +148,11 @@ jobs:
          example: ${{ inputs.example }}
          hardware: ${{ inputs.hardware }}
          test_case: ${{ matrix.test_case }}
+          use_model_cache: ${{ inputs.use_model_cache }}
        run: |
          cd ${{ github.workspace }}/$example/tests
          if [[ "$IMAGE_REPO" == "" ]]; then export IMAGE_REPO="${OPEA_IMAGE_REPO}opea"; fi
+          if [[ "$use_model_cache" == "true" ]]; then export model_cache="/data2/hf_model"; fi
          if [ -f ${test_case} ]; then timeout 30m bash ${test_case}; else echo "Test script {${test_case}} not found, skip test!"; fi

      - name: Clean up container after test
--- a/.github/workflows/manual-example-workflow.yml
+++ b/.github/workflows/manual-example-workflow.yml
@@ -20,11 +20,11 @@ on:
        description: "Tag to apply to images"
        required: true
        type: string
-      deploy_gmc:
-        default: false
-        description: 'Whether to deploy gmc'
-        required: true
-        type: boolean
+      # deploy_gmc:
+      #   default: false
+      #   description: 'Whether to deploy gmc'
+      #   required: true
+      #   type: boolean
      build:
        default: true
        description: 'Build test required images for Examples'
@@ -40,11 +40,11 @@ on:
        description: 'Test examples with helm charts'
        required: false
        type: boolean
-      test_gmc:
-        default: false
-        description: 'Test examples with gmc'
-        required: false
-        type: boolean
+      # test_gmc:
+      #   default: false
+      #   description: 'Test examples with gmc'
+      #   required: false
+      #   type: boolean
      opea_branch:
        default: "main"
        description: 'OPEA branch for image build'
@@ -54,7 +54,12 @@ on:
        default: false
        description: "inject commit to docker images true or false"
        required: false
-        type: string
+        type: boolean
+      use_model_cache:
+        default: false
+        description: "use model cache true or false"
+        required: false
+        type: boolean

 permissions: read-all
 jobs:
@@ -76,7 +81,8 @@ jobs:

  build-deploy-gmc:
    needs: [get-test-matrix]
-    if: ${{ fromJSON(inputs.deploy_gmc) }}
+    if: false
+    #${{ fromJSON(inputs.deploy_gmc) }}
    strategy:
      matrix:
        node: ${{ fromJson(needs.get-test-matrix.outputs.nodes) }}
@@ -89,7 +95,7 @@ jobs:
    secrets: inherit

  run-examples:
-    needs: [get-test-matrix, build-deploy-gmc]
+    needs: [get-test-matrix]   #[get-test-matrix, build-deploy-gmc]
    if: always()
    strategy:
      matrix:
@@ -104,7 +110,8 @@ jobs:
      build: ${{ fromJSON(inputs.build) }}
      test_compose: ${{ fromJSON(inputs.test_compose) }}
      test_helmchart: ${{ fromJSON(inputs.test_helmchart) }}
-      test_gmc: ${{ fromJSON(inputs.test_gmc) }}
+      # test_gmc: ${{ fromJSON(inputs.test_gmc) }}
      opea_branch: ${{ inputs.opea_branch }}
      inject_commit: ${{ inputs.inject_commit }}
+      use_model_cache: ${{ inputs.use_model_cache }}
    secrets: inherit
--- a/.github/workflows/pr-docker-compose-e2e.yml
+++ b/.github/workflows/pr-docker-compose-e2e.yml
@@ -42,5 +42,6 @@ jobs:
      tag: "ci"
      example: ${{ matrix.example }}
      hardware: ${{ matrix.hardware }}
+      use_model_cache: true
      diff_excluded_files: '\.github|\.md|\.txt|kubernetes|gmc|assets|benchmark'
    secrets: inherit
--- a/AudioQnA/docker_compose/intel/cpu/xeon/compose.yaml
+++ b/AudioQnA/docker_compose/intel/cpu/xeon/compose.yaml
@@ -30,7 +30,7 @@ services:
    ports:
      - "3006:80"
    volumes:
-      - "./data:/data"
+      - "${MODEL_CACHE}:/data"
    shm_size: 1g
    environment:
      no_proxy: ${no_proxy}
--- a/AudioQnA/docker_compose/intel/cpu/xeon/compose_multilang.yaml
+++ b/AudioQnA/docker_compose/intel/cpu/xeon/compose_multilang.yaml
@@ -31,7 +31,7 @@ services:
    ports:
      - "3006:80"
    volumes:
-      - "./data:/data"
+      - "${MODEL_CACHE}:/data"
    shm_size: 1g
    environment:
      no_proxy: ${no_proxy}
--- a/AudioQnA/docker_compose/intel/hpu/gaudi/compose.yaml
+++ b/AudioQnA/docker_compose/intel/hpu/gaudi/compose.yaml
@@ -40,7 +40,7 @@ services:
    ports:
      - "3006:80"
    volumes:
-      - "./data:/data"
+      - "${MODEL_CACHE}:/data"
    environment:
      no_proxy: ${no_proxy}
      http_proxy: ${http_proxy}
--- a/AudioQnA/tests/test_compose_on_gaudi.sh
+++ b/AudioQnA/tests/test_compose_on_gaudi.sh
@@ -9,6 +9,7 @@ echo "REGISTRY=IMAGE_REPO=${IMAGE_REPO}"
 echo "TAG=IMAGE_TAG=${IMAGE_TAG}"
 export REGISTRY=${IMAGE_REPO}
 export TAG=${IMAGE_TAG}
+export MODEL_CACHE=${model_cache:-"./data"}

 WORKPATH=$(dirname "$PWD")
 LOG_PATH="$WORKPATH/tests"
--- a/AudioQnA/tests/test_compose_on_xeon.sh
+++ b/AudioQnA/tests/test_compose_on_xeon.sh
@@ -9,6 +9,7 @@ echo "REGISTRY=IMAGE_REPO=${IMAGE_REPO}"
 echo "TAG=IMAGE_TAG=${IMAGE_TAG}"
 export REGISTRY=${IMAGE_REPO}
 export TAG=${IMAGE_TAG}
+export MODEL_CACHE=${model_cache:-"./data"}

 WORKPATH=$(dirname "$PWD")
 LOG_PATH="$WORKPATH/tests"
--- a/DBQnA/docker_compose/intel/cpu/xeon/compose.yaml
+++ b/DBQnA/docker_compose/intel/cpu/xeon/compose.yaml
@@ -8,7 +8,7 @@ services:
    ports:
      - "8008:80"
    volumes:
-      - "./data:/data"
+      - "${MODEL_CACHE}:/data"
    environment:
      no_proxy: ${no_proxy}
      http_proxy: ${http_proxy}
--- a/DBQnA/tests/test_compose_on_xeon.sh
+++ b/DBQnA/tests/test_compose_on_xeon.sh
@@ -10,6 +10,7 @@ echo "REGISTRY=IMAGE_REPO=${IMAGE_REPO}"
 echo "TAG=IMAGE_TAG=${IMAGE_TAG}"
 export REGISTRY=${IMAGE_REPO}
 export TAG=${IMAGE_TAG}
+export MODEL_CACHE=${model_cache:-"./data"}

 WORKPATH=$(dirname "$PWD")
 LOG_PATH="$WORKPATH/tests"
--- a/DocIndexRetriever/docker_compose/intel/cpu/xeon/compose.yaml
+++ b/DocIndexRetriever/docker_compose/intel/cpu/xeon/compose.yaml
@@ -38,7 +38,7 @@ services:
    ports:
      - "6006:80"
    volumes:
-      - "./data:/data"
+      - "${MODEL_CACHE}:/data"
    shm_size: 1g
    environment:
      no_proxy: ${no_proxy}
@@ -96,7 +96,7 @@ services:
    ports:
      - "8808:80"
    volumes:
-      - "./data:/data"
+      - "${MODEL_CACHE}:/data"
    shm_size: 1g
    environment:
      no_proxy: ${no_proxy}
--- a/DocIndexRetriever/docker_compose/intel/cpu/xeon/compose_without_rerank.yaml
+++ b/DocIndexRetriever/docker_compose/intel/cpu/xeon/compose_without_rerank.yaml
@@ -36,7 +36,7 @@ services:
    ports:
      - "6006:80"
    volumes:
-      - "/mnt/models:/data"
+      - "${MODEL_CACHE:-/mnt/models}:/data"
    shm_size: 1g
    environment:
      no_proxy: ${no_proxy}
--- a/DocIndexRetriever/docker_compose/intel/hpu/gaudi/compose.yaml
+++ b/DocIndexRetriever/docker_compose/intel/hpu/gaudi/compose.yaml
@@ -34,7 +34,7 @@ services:
    ports:
      - "8090:80"
    volumes:
-      - "./data:/data"
+      - "${MODEL_CACHE}:/data"
    runtime: habana
    cap_add:
      - SYS_NICE
@@ -95,7 +95,7 @@ services:
    ports:
      - "8808:80"
    volumes:
-      - "./data:/data"
+      - "${MODEL_CACHE}:/data"
    shm_size: 1g
    environment:
      no_proxy: ${no_proxy}
--- a/DocIndexRetriever/tests/test_compose_on_gaudi.sh
+++ b/DocIndexRetriever/tests/test_compose_on_gaudi.sh
@@ -9,6 +9,7 @@ echo "REGISTRY=IMAGE_REPO=${IMAGE_REPO}"
 echo "TAG=IMAGE_TAG=${IMAGE_TAG}"
 export REGISTRY=${IMAGE_REPO}
 export TAG=${IMAGE_TAG}
+export MODEL_CACHE=${model_cache:-"./data"}

 WORKPATH=$(dirname "$PWD")
 LOG_PATH="$WORKPATH/tests"
--- a/DocIndexRetriever/tests/test_compose_on_xeon.sh
+++ b/DocIndexRetriever/tests/test_compose_on_xeon.sh
@@ -9,6 +9,7 @@ echo "REGISTRY=IMAGE_REPO=${IMAGE_REPO}"
 echo "TAG=IMAGE_TAG=${IMAGE_TAG}"
 export REGISTRY=${IMAGE_REPO}
 export TAG=${IMAGE_TAG}
+export MODEL_CACHE=${model_cache:-"./data"}

 WORKPATH=$(dirname "$PWD")
 LOG_PATH="$WORKPATH/tests"
--- a/DocIndexRetriever/tests/test_compose_without_rerank_on_xeon.sh
+++ b/DocIndexRetriever/tests/test_compose_without_rerank_on_xeon.sh
@@ -9,6 +9,7 @@ echo "REGISTRY=IMAGE_REPO=${IMAGE_REPO}"
 echo "TAG=IMAGE_TAG=${IMAGE_TAG}"
 export REGISTRY=${IMAGE_REPO}
 export TAG=${IMAGE_TAG}
+export MODEL_CACHE=${model_cache:-"./data"}

 WORKPATH=$(dirname "$PWD")
 LOG_PATH="$WORKPATH/tests"
--- a/FaqGen/docker_compose/intel/cpu/xeon/compose.yaml
+++ b/FaqGen/docker_compose/intel/cpu/xeon/compose.yaml
@@ -8,7 +8,7 @@ services:
    ports:
      - ${LLM_ENDPOINT_PORT:-8008}:80
    volumes:
-      - "./data:/data"
+      - "${MODEL_CACHE}:/data"
    shm_size: 1g
    environment:
      no_proxy: ${no_proxy}
--- a/FaqGen/tests/test_compose_on_gaudi.sh
+++ b/FaqGen/tests/test_compose_on_gaudi.sh
@@ -13,7 +13,7 @@ export TAG=${IMAGE_TAG}
 WORKPATH=$(dirname "$PWD")
 LOG_PATH="$WORKPATH/tests"
 ip_address=$(hostname -I | awk '{print $1}')
-export DATA_PATH="/data/cache"
+export DATA_PATH=${model_cache:-"/data/cache"}

 function build_docker_images() {
    opea_branch=${opea_branch:-"main"}
--- a/FaqGen/tests/test_compose_on_xeon.sh
+++ b/FaqGen/tests/test_compose_on_xeon.sh
@@ -9,6 +9,7 @@ echo "REGISTRY=IMAGE_REPO=${IMAGE_REPO}"
 echo "TAG=IMAGE_TAG=${IMAGE_TAG}"
 export REGISTRY=${IMAGE_REPO}
 export TAG=${IMAGE_TAG}
+export MODEL_CACHE=${model_cache:-"./data"}

 WORKPATH=$(dirname "$PWD")
 LOG_PATH="$WORKPATH/tests"
--- a/Translation/docker_compose/intel/cpu/xeon/compose.yaml
+++ b/Translation/docker_compose/intel/cpu/xeon/compose.yaml
@@ -21,7 +21,7 @@ services:
      timeout: 10s
      retries: 100
    volumes:
-      - "./data:/data"
+      - "${MODEL_CACHE}:/data"
    shm_size: 1g
    command: --model-id ${LLM_MODEL_ID} --cuda-graphs 0
  llm:
--- a/Translation/docker_compose/intel/hpu/gaudi/compose.yaml
+++ b/Translation/docker_compose/intel/hpu/gaudi/compose.yaml
@@ -30,7 +30,7 @@ services:
      - SYS_NICE
    ipc: host
    volumes:
-      - "./data:/data"
+      - "${MODEL_CACHE}:/data"
    command: --model-id ${LLM_MODEL_ID} --max-input-length 1024 --max-total-tokens 2048
  llm:
    image: ${REGISTRY:-opea}/llm-textgen:${TAG:-latest}
--- a/Translation/tests/test_compose_on_gaudi.sh
+++ b/Translation/tests/test_compose_on_gaudi.sh
@@ -9,6 +9,7 @@ echo "REGISTRY=IMAGE_REPO=${IMAGE_REPO}"
 echo "TAG=IMAGE_TAG=${IMAGE_TAG}"
 export REGISTRY=${IMAGE_REPO}
 export TAG=${IMAGE_TAG}
+export MODEL_CACHE=${model_cache:-"./data"}

 WORKPATH=$(dirname "$PWD")
 LOG_PATH="$WORKPATH/tests"
--- a/Translation/tests/test_compose_on_xeon.sh
+++ b/Translation/tests/test_compose_on_xeon.sh
@@ -9,6 +9,7 @@ echo "REGISTRY=IMAGE_REPO=${IMAGE_REPO}"
 echo "TAG=IMAGE_TAG=${IMAGE_TAG}"
 export REGISTRY=${IMAGE_REPO}
 export TAG=${IMAGE_TAG}
+export MODEL_CACHE=${model_cache:-"./data"}

 WORKPATH=$(dirname "$PWD")
 LOG_PATH="$WORKPATH/tests"
--- a/VisualQnA/docker_compose/intel/cpu/xeon/compose.yaml
+++ b/VisualQnA/docker_compose/intel/cpu/xeon/compose.yaml
@@ -8,7 +8,7 @@ services:
    ports:
      - "8399:80"
    volumes:
-      - "./data:/data"
+      - "${MODEL_CACHE}:/data"
    shm_size: 1g
    environment:
      no_proxy: ${no_proxy}
--- a/VisualQnA/docker_compose/intel/hpu/gaudi/compose.yaml
+++ b/VisualQnA/docker_compose/intel/hpu/gaudi/compose.yaml
@@ -8,7 +8,7 @@ services:
    ports:
      - "8399:80"
    volumes:
-      - "./data:/data"
+      - "${MODEL_CACHE}:/data"
    environment:
      no_proxy: ${no_proxy}
      http_proxy: ${http_proxy}
--- a/VisualQnA/tests/test_compose_on_gaudi.sh
+++ b/VisualQnA/tests/test_compose_on_gaudi.sh
@@ -9,6 +9,7 @@ echo "REGISTRY=IMAGE_REPO=${IMAGE_REPO}"
 echo "TAG=IMAGE_TAG=${IMAGE_TAG}"
 export REGISTRY=${IMAGE_REPO}
 export TAG=${IMAGE_TAG}
+export MODEL_CACHE=${model_cache:-"./data"}

 WORKPATH=$(dirname "$PWD")
 LOG_PATH="$WORKPATH/tests"
--- a/VisualQnA/tests/test_compose_on_xeon.sh
+++ b/VisualQnA/tests/test_compose_on_xeon.sh
@@ -9,6 +9,7 @@ echo "REGISTRY=IMAGE_REPO=${IMAGE_REPO}"
 echo "TAG=IMAGE_TAG=${IMAGE_TAG}"
 export REGISTRY=${IMAGE_REPO}
 export TAG=${IMAGE_TAG}
+export MODEL_CACHE=${model_cache:-"./data"}

 WORKPATH=$(dirname "$PWD")
 LOG_PATH="$WORKPATH/tests"