Enable Intel ARC gpu test for vllm openvino.

Signed-off-by: senhui2intel <[email protected]>

tests/llms/test_llms_text-generation_vllm_langchain_openvino_on_intel_arc.sh

@@ -0,0 +1,150 @@
+#!/bin/bash
+# Copyright (C) 2024 Intel Corporation
+# SPDX-License-Identifier: Apache-2.0
+
+set -x
+
+WORKPATH="$( cd "$( dirname "$0" )" && pwd )"
+
+# Define variables
+port=5033
+RENDER_GROUP_ID=110
+HF_MODEL_FIR=$HOME/hf_model
+DOCKER_IMAGE="vllm-openvino:comps"
+CONTAINER_NAME="test-comps-vllm-openvino-container"
+
+function build_container() {
+    cd $WORKPATH
+    git clone https://github.com/vllm-project/vllm.git vllm-openvino
+    cd ./vllm-openvino
+
+    git reset --hard 067e77f9a87c3466fce41c8fe8710fddc69ec26c # resolve circular import issue
+
+    # Add ARC drive to dockerfile
+    sed -i '9r /dev/stdin' Dockerfile.openvino << EndOfMessage
+RUN apt-get install -y gpg-agent wget
+RUN wget -qO - https://repositories.intel.com/gpu/intel-graphics.key | gpg --yes --dearmor --output /usr/share/keyrings/intel-graphics.gpg && \\
+    echo "deb [arch=amd64 signed-by=/usr/share/keyrings/intel-graphics.gpg] https://repositories.intel.com/gpu/ubuntu jammy/lts/2350 unified" | \\
+        tee /etc/apt/sources.list.d/intel-gpu-jammy.list && \\
+    apt-get update -y && \\
+    apt-get install -y \\
+        intel-opencl-icd intel-level-zero-gpu level-zero \\
+        intel-media-va-driver-non-free libmfx1 libmfxgen1 libvpl2 \\
+        libegl-mesa0 libegl1-mesa libegl1-mesa-dev libgbm1 libgl1-mesa-dev libgl1-mesa-dri \\
+        libglapi-mesa libgles2-mesa-dev libglx-mesa0 libigdgmm12 libxatracker2 mesa-va-drivers \\
+        mesa-vdpau-drivers mesa-vulkan-drivers va-driver-all vainfo hwinfo clinfo
+EndOfMessage
+
+    docker build --no-cache -t $DOCKER_IMAGE \
+      -f Dockerfile.openvino \
+      . \
+      --build-arg https_proxy=$https_proxy \
+      --build-arg http_proxy=$http_proxy
+    if [ $? -ne 0 ]; then
+        echo "vllm-openvino built fail"
+        exit 1
+    else
+        echo "vllm-openvino built successful"
+    fi
+    cd $WORKPATH
+    rm -rf vllm-openvino
+}
+
+# Function to start Docker container
+start_container() {
+
+    docker run -d --rm --name=$CONTAINER_NAME \
+      -p $port:$port \
+      --ipc=host \
+      -e HTTPS_PROXY=$https_proxy \
+      -e HTTP_PROXY=$https_proxy \
+      -v $HF_MODEL_FIR:/hf_model \
+      --device=/dev/dri:/dev/dri \
+      --group-add $RENDER_GROUP_ID
+      vllm-openvino:comps /bin/bash -c "\
+        export VLLM_CPU_KVCACHE_SPACE=8 && \
+        export VLLM_OPENVINO_DEVICE=GPU && \
+        export VLLM_OPENVINO_ENABLE_QUANTIZED_WEIGHTS=ON && \
+        python3 -m vllm.entrypoints.openai.api_server \
+          --model /hf_model/neural-chat-7b-v3-3 \
+          --host 0.0.0.0 \
+          --port $port"
+
+    # check whether service is fully ready
+    n=0
+    until [[ "$n" -ge 300 ]]; do
+        docker logs $CONTAINER_NAME > /tmp/$CONTAINER_NAME.log 2>&1
+        n=$((n+1))
+        if grep -q "Uvicorn running on" /tmp/$CONTAINER_NAME.log; then
+            break
+        fi
+        sleep 3s
+    done
+
+}
+
+# Cleanup Function
+cleanup() {
+    # Stop and remove Docker container and images
+    cid=$(docker ps -aq --filter "name=$CONTAINER_NAME")
+        if [[ ! -z "$cid" ]]; then docker stop $cid || docker rm $cid && sleep 1s; fi
+    docker rmi -f $DOCKER_IMAGE
+    rm /tmp/$CONTAINER_NAME.log
+}
+
+# Function to test API endpoint
+function test_api_endpoint {
+    local endpoint="$1"
+    local expected_status="$2"
+
+    # Make the HTTP request
+    if test "$1" = "v1/completions"
+    then
+        local response=$(curl "http://localhost:$port/$endpoint" \
+          -H "Content-Type: application/json" \
+          -d '{
+                "model": "/hf_model/neural-chat-7b-v3-3",
+                "prompt": "What is the key advantage of Openvino framework",
+                "max_tokens": 300,
+                "temperature": 0.7
+              }' \
+          --write-out '%{http_code}' \
+          --silent \
+          --output /dev/null)
+    else
+        local response=$(curl "http://localhost:$port/$endpoint" \
+          --write-out '%{http_code}' \
+          --silent \
+          --output /dev/null)
+    fi
+
+    # Assert the response status code
+    if [[ "$response" -eq "$expected_status" ]]; then
+        echo "PASS: $endpoint returned expected status code: $expected_status"
+    else
+        echo "FAIL: $endpoint returned unexpected status code: $response (expected: $expected_status)"
+        docker logs $CONTAINER_NAME
+        exit 1
+    fi
+}
+
+# Main function
+main() {
+
+    # use local image to skip slow network connection
+    # build_container
+    start_container
+
+    # Sleep to allow the container to start up fully
+    sleep 10
+    # Test the /v1/models API
+    test_api_endpoint "v1/models" 200
+
+    # Test the /v1/completions API
+    test_api_endpoint "v1/completions" 200
+
+    cleanup
+}
+
+# Call main function
+main