refactor: move smoke tests into a seperate shell script

also add shell linter to pre-commit config

Assisted-by: coderabbitai
Signed-off-by: Nathan Weinberg <nweinber@redhat.com>

Author: nathan-weinberg

.github/workflows/redhat-distro-container.yml

@@ -55,49 +55,12 @@ jobs:
         id: vllm
         uses: ./.github/actions/setup-vllm
 
-      - name: Test image
-        id: test
-        run: |
-          set -euo pipefail
-          # Start llama stack
-          CID="$(docker run -d --pull=never \
-            -p 8321:8321 \
-            --env INFERENCE_MODEL=meta-llama/Llama-3.2-1B-Instruct \
-            --env TRUSTYAI_LMEVAL_USE_K8S=False \
-            --name llama-stack \
-            ${{ env.IMAGE_NAME }}:${{ github.sha }})"
-          trap 'docker rm -f "$CID" >/dev/null 2>&1 || true' EXIT
-          echo "Started Llama Stack container with CID: $CID"
-
-          echo "Waiting for Llama Stack server..."
-          for i in {1..60}; do
-            echo "Attempt $i to connect to Llama Stack..."
-            if curl -fsS --max-time 2 http://127.0.0.1:8321/v1/health | grep -q '"status":"OK"'; then
-              echo "Llama Stack server is up and serving :)"
-              if curl -fsS --max-time 4 http://127.0.0.1:8321/v1/models | grep -q 'meta-llama/Llama-3.2-1B-Instruct'; then
-                echo "meta-llama/Llama-3.2-1B-Instruct model was found :)"
-                if curl -fsS --max-time 6 http://127.0.0.1:8321/v1/openai/v1/chat/completions -H "Content-Type: application/json" -d "{\"model\": \"meta-llama/Llama-3.2-1B-Instruct\",\"messages\": [{\"role\": \"user\", \"content\": \"What color is grass?\"}], \"max_tokens\": 10, \"temperature\": 0.0}" | grep -q 'green'; then
-                  echo "Inference is working :)"
-                  exit 0
-                else
-                  echo "Inference is not working :("
-                  echo "Container logs:"
-                  docker logs "$CID" || true
-                  exit 1
-                fi
-              else
-                echo "meta-llama/Llama-3.2-1B-Instruct model was not found :("
-                echo "Container logs:"
-                docker logs "$CID" || true
-                exit 1
-              fi
-            fi
-            sleep 1
-          done
-          echo "Llama Stack server failed to start :("
-          echo "Container logs:"
-          docker logs "$CID" || true
-          exit 1
+      - name: Smoke test image
+        id: smoke-test
+        shell: bash
+        env:
+          INFERENCE_MODEL: meta-llama/Llama-3.2-1B-Instruct
+        run: ./tests/smoke.sh
 
       - name: Log in to Quay.io
         id: login

.pre-commit-config.yaml

@@ -40,6 +40,11 @@ repos:
     hooks:
       - id: actionlint
 
+-   repo: https://github.com/koalaman/shellcheck-precommit
+    rev: v0.11.0
+    hooks:
+    -   id: shellcheck
+
 -   repo: local
     hooks:
       - id: pkg-gen

tests/smoke.sh

@@ -0,0 +1,72 @@
+#!/bin/bash
+
+set -uo pipefail
+
+function start_and_wait_for_llama_stack_container {
+  # Start llama stack
+  docker run \
+    -d \
+    --pull=never \
+    -p 8321:8321 \
+    --env INFERENCE_MODEL="$INFERENCE_MODEL" \
+    --env TRUSTYAI_LMEVAL_USE_K8S=False \
+    --name llama-stack \
+    "$IMAGE_NAME:$GITHUB_SHA"
+  echo "Started Llama Stack container..."
+
+  # Wait for llama stack to be ready by doing a health check
+  echo "Waiting for Llama Stack server..."
+  for i in {1..60}; do
+    echo "Attempt $i to connect to Llama Stack..."
+    resp=$(curl -fsS --max-time 2 http://127.0.0.1:8321/v1/health)
+    if [ "$resp" == '{"status":"OK"}' ]; then
+      echo "Llama Stack server is up!"
+      return
+    fi
+    sleep 1
+  done
+  echo "Llama Stack server failed to start :("
+  echo "Container logs:"
+  docker logs llama-stack || true
+  exit 1
+}
+
+function test_model_list {
+  echo "===> Looking for model $INFERENCE_MODEL..."
+  resp=$(curl -fsS --max-time 4 http://127.0.0.1:8321/v1/models)
+  if echo "$resp" | grep -q "$INFERENCE_MODEL"; then
+    echo "Model $INFERENCE_MODEL was found :)"
+    return
+  else
+    echo "Model $INFERENCE_MODEL was not found :("
+    echo "Container logs:"
+    docker logs llama-stack || true
+    exit 1
+  fi
+}
+
+function test_model_openai_inference {
+  echo "===> Attempting to chat with model $INFERENCE_MODEL..."
+  resp=$(curl -fsS --max-time 6 http://127.0.0.1:8321/v1/openai/v1/chat/completions -H "Content-Type: application/json" -d "{\"model\": \"$INFERENCE_MODEL\",\"messages\": [{\"role\": \"user\", \"content\": \"What color is grass?\"}], \"max_tokens\": 10, \"temperature\": 0.0}")
+  if echo "$resp" | grep -q "green"; then
+    echo "===> Inference is working :)"
+    return
+  else
+    echo "===> Inference is not working :("
+    echo "Container logs:"
+    docker logs llama-stack || true
+    exit 1
+  fi
+}
+
+main() {
+  echo "===> Starting smoke test..."
+  start_and_wait_for_llama_stack_container
+  test_model_list
+  test_model_openai_inference
+  echo "===> Smoke test completed successfully!"
+}
+
+trap 'docker rm -f -v llama-stack >/dev/null 2>&1 || true' EXIT
+main "$@"
+exit 0