feat(ci): add support for running vision inference tests

.github/actions/run-integration-tests/action.yml

@@ -14,6 +14,10 @@ inputs:
   inference-mode:
     description: 'Inference mode: "record" or "replay"'
     required: true
+  run-vision-tests:
+    description: 'Run vision tests: "true" or "false"'
+    required: false
+    default: 'false'
 
 outputs:
   logs-path:
@@ -46,6 +50,22 @@ runs:
           EXCLUDE_TESTS="${EXCLUDE_TESTS} or test_inference_store_tool_calls"
         fi
 
+        if [ "${{ inputs.run-vision-tests }}" == "true" ]; then
+          if uv run pytest -s -v tests/integration/inference/test_vision_inference.py --stack-config=${stack_config} \
+            -k "not( ${EXCLUDE_TESTS} )" \
+            --vision-model=ollama/llama3.2-vision:11b \
+            --embedding-model=sentence-transformers/all-MiniLM-L6-v2 \
+            --color=yes ${EXTRA_PARAMS} \
+            --capture=tee-sys | tee pytest-${{ inputs.inference-mode }}-vision.log; then
+            echo "✅ Tests completed for vision"
+          else
+            echo "❌ Tests failed for vision"
+            exit 1
+          fi
+
+          exit 0
+        fi
+
         TEST_TYPES='${{ inputs.test-types }}'
         echo "Test types to run: $TEST_TYPES"
 

.github/actions/setup-vision-ollama/action.yml

@@ -0,0 +1,11 @@
+name: Setup Ollama for Vision Tests
+description: Start Ollama with vision model
+runs:
+  using: "composite"
+  steps:
+    - name: Start Ollama
+      shell: bash
+      run: |
+        docker run -d --name ollama -p 11434:11434 docker.io/llamastack/ollama-with-vision-model
+        echo "Verifying Ollama status..."
+        timeout 30 bash -c 'while ! curl -s -L http://127.0.0.1:11434; do sleep 1 && echo "."; done'