update ut.

Author: lkk12014402

FinanceAgent/docker_compose/intel/hpu/gaudi/compose.yaml

@@ -60,6 +60,7 @@ services:
       http_proxy: ${http_proxy}
       https_proxy: ${https_proxy}
       FINNHUB_API_KEY: ${FINNHUB_API_KEY}
+      FINANCIAL_DATASETS_API_KEY: ${FINANCIAL_DATASETS_API_KEY}
       port: 9096
 
   supervisor-react-agent:

FinanceAgent/docker_compose/intel/hpu/gaudi/launch_agents.sh

@@ -30,6 +30,7 @@ export MAX_TOTAL_TOKENS=4096
 export DocSum_COMPONENT_NAME="OpeaDocSumvLLM"
 export DOCSUM_ENDPOINT="http://${ip_address}:9000/v1/docsum"
 
-
+export FINNHUB_API_KEY=${FINNHUB_API_KEY}
+export FINANCIAL_DATASETS_API_KEY=${FINANCIAL_DATASETS_API_KEY}
 
 docker compose -f compose.yaml up -d

FinanceAgent/prompts/research_prompt.py

@@ -26,38 +26,23 @@
 
 For writing a comprehensive analysis financial research report, you can use all the tools provided to retrieve information available for the company.
 
-**Procedures:**
+**Pay attention to the followings:**
 1. Explicitly explain your working plan before you kick off.
 2. Read the question carefully. Firstly You need get accurate `start_date` and `end_date` value, because most tools need the 2 values like company news, financials. You can get `end_date` with `get_current_date` tool if user doesn't provide. And you can infer `start_date` with `end_date` using the rule `start_date is one year earlier than end_date` if user doesn't provide.
-3. Analyze Key Financial Statements:
-    - **Balance Sheet Analysis**: Use the `analyze_balance_sheet` tool to obtain insights into the company's financial stability and liquidity.
-    - **Income Statement Analysis**: Employ the `analyze_income_stmt` tool to assess the company's profitability over the fiscal year.
-    - **Cash Flow Statement Analysis**: Apply the `analyze_cash_flow` tool to evaluate the cash inflows and outflows, providing a clear view of the company's operational efficiency.
-4. Perform Additional Analyses:
-    - **Business Highlights**: Use the `get_company_news` tool to summarize the company's operational achievements and strategic milestones over the year.
-    - **Company Description**: Utilize the `get_company_profile` tool to provide a detailed overview of the company's business model, industry positioning, and strategic initiatives.
-    - **Risk Assessment**: Employ all the tool output financial information to identify and summarize the top three risks facing the company.
-5. Share Performance: analysis the stock performance using the `get_share_performance` tool.
-6. Generate Report: Finally, compile all the analyzed data, insights, and visuals into a comprehensive financial report, which contains the following paragraphs:
-    - income summarization
-    - market position
-    - business overview
-    - risk assessment
-    - competitors analysis
-    - share performance analysis
-7. All the paragraphs should combine between 400 and 450 words.
-8. Read the execution history if any to understand the tools that have been called and the information that has been gathered.
-9. Reason about the information gathered so far and decide if you can answer the question or if you need to call more tools.
-10. Do not repeat your thinking process!!! Only gather all information when you generate report.
+3. Use tools one by one for clarity, especially when asking for instructions.
+4. Provide stock performace, because the financial report is used for stock investment analysis.
+5. Read the execution history if any to understand the tools that have been called and the information that has been gathered.
+6. Reason about the information gathered so far and decide if you can answer the question or if you need to call more tools.
 
 **Output format:**
-You should output your thought process. Finish thinking first. Output tool calls or your answer at the end.
-When making tool calls, you should use the following format:
+You should output your thought process:
+
+When need tool calls, you should use the following format:
 TOOL CALL: {{"tool": "tool1", "args": {{"arg1": "value1", "arg2": "value2", ...}}}}
 TOOL CALL: {{"tool": "tool2", "args": {{"arg1": "value1", "arg2": "value2", ...}}}}
 
-If you can generate the financial, provide the report in the following format:
-FINAL ANSWER: {{"answer": "your financial research report here"}}
+If you have enough financial data, provide the financial report in the following format:
+FINAL ANSWER: {{"answer": "compile all the analyzed data and insights into a comprehensive financial report, which contains the following paragraphs: income summarization, market position, business overview, risk assessment, competitors analysis, share performance analysis."}}
 
 Follow these guidelines when formulating your answer:
 1. If the question contains a false premise or assumption, answer “invalid question”.

FinanceAgent/tests/test.py

@@ -29,7 +29,10 @@ def process_request(url, query, is_stream=False):
 
 def test_worker_agent(args):
     url = f"http://{args.ip_addr}:{args.ext_port}/v1/chat/completions"
-    query = {"role": "user", "messages": args.prompt, "stream": "false"}
+    if args.tool_choice is None:
+        query = {"role": "user", "messages": args.prompt, "stream": "false"}
+    else:
+        query = {"role": "user", "messages": args.prompt, "stream": "false", "tool_choice": args.tool_choice}
     ret = process_request(url, query)
     print("Response: ", ret)
 
@@ -79,6 +82,7 @@ def test_supervisor_agent_single_turn(args):
     parser.add_argument("--prompt", type=str, help="prompt message")
     parser.add_argument("--agent_role", type=str, default="supervisor", help="supervisor or worker")
     parser.add_argument("--multi-turn", action="store_true", help="multi-turn conversation")
+    parser.add_argument('--tool_choice', nargs='+', help='limit tools')
     args, _ = parser.parse_known_args()
 
     print(args)

FinanceAgent/tests/test_compose_on_gaudi.sh

@@ -59,7 +59,8 @@ function build_vllm_docker_image() {
         git clone https://github.com/HabanaAI/vllm-fork.git
     fi
     cd ./vllm-fork
-    VLLM_VER=$(git describe --tags "$(git rev-list --tags --max-count=1)")
+    # VLLM_VER=$(git describe --tags "$(git rev-list --tags --max-count=1)")
+    VLLM_VER=v0.6.6.post1+Gaudi-1.20.0
     git checkout ${VLLM_VER} &> /dev/null
     docker build --no-cache -f Dockerfile.hpu -t $vllm_image --shm-size=128g . --build-arg https_proxy=$https_proxy --build-arg http_proxy=$http_proxy
     if [ $? -ne 0 ]; then
@@ -75,11 +76,11 @@ function start_vllm_service_70B() {
     echo "token is ${HF_TOKEN}"
     echo "start vllm gaudi service"
     echo "**************model is $model**************"
-    docker run -d --runtime=habana --rm --name "vllm-gaudi-server" -e HABANA_VISIBLE_DEVICES=0,1,2,3 -p $vllm_port:8000 -v $vllm_volume:/data -e HF_TOKEN=$HF_TOKEN -e HUGGING_FACE_HUB_TOKEN=$HF_TOKEN -e HF_HOME=/data -e OMPI_MCA_btl_vader_single_copy_mechanism=none -e PT_HPU_ENABLE_LAZY_COLLECTIVES=true -e http_proxy=$http_proxy -e https_proxy=$https_proxy -e no_proxy=$no_proxy -e VLLM_SKIP_WARMUP=true --cap-add=sys_nice --ipc=host $vllm_image --model ${model} --max-seq-len-to-capture 16384 --tensor-parallel-size 4
-    sleep 5s
+    docker run -d --runtime=habana --rm --name "vllm-gaudi-server" -e HABANA_VISIBLE_DEVICES=all -p $vllm_port:8000 -v $vllm_volume:/data -e HF_TOKEN=$HF_TOKEN -e HUGGING_FACE_HUB_TOKEN=$HF_TOKEN -e HF_HOME=/data -e OMPI_MCA_btl_vader_single_copy_mechanism=none -e PT_HPU_ENABLE_LAZY_COLLECTIVES=true -e http_proxy=$http_proxy -e https_proxy=$https_proxy -e no_proxy=$no_proxy -e VLLM_SKIP_WARMUP=true --cap-add=sys_nice --ipc=host $vllm_image --model ${model} --max-seq-len-to-capture 16384 --tensor-parallel-size 4
+    sleep 10s
     echo "Waiting vllm gaudi ready"
     n=0
-    until [[ "$n" -ge 100 ]] || [[ $ready == true ]]; do
+    until [[ "$n" -ge 200 ]] || [[ $ready == true ]]; do
         docker logs vllm-gaudi-server &> ${LOG_PATH}/vllm-gaudi-service.log
         n=$((n+1))
         if grep -q "Uvicorn running on" ${LOG_PATH}/vllm-gaudi-service.log; then
@@ -89,9 +90,9 @@ function start_vllm_service_70B() {
             echo "container vllm-gaudi-server not found"
             exit 1
         fi
-        sleep 5s
+        sleep 10s
     done
-    sleep 5s
+    sleep 10s
     echo "Service started successfully"
 }
 
@@ -181,7 +182,7 @@ function validate_agent_service() {
     echo "======================Testing worker research agent======================"
     export agent_port="9096"
     prompt="generate NVDA financial research report"
-    local CONTENT=$(python3 $WORKDIR/GenAIExamples/AgentQnA/tests/test.py --prompt "$prompt" --agent_role "worker" --ext_port $agent_port)
+    local CONTENT=$(python3 $WORKDIR/GenAIExamples/AgentQnA/tests/test.py --prompt "$prompt" --agent_role "worker" --ext_port $agent_port --tool_choice "get_current_date" --tool_choice "get_share_performance")
     local EXIT_CODE=$(validate "$CONTENT" "NVDA" "research-agent-endpoint")
     echo $CONTENT
     echo $EXIT_CODE
@@ -214,6 +215,7 @@ function validate_agent_service() {
     #     docker logs react-agent-endpoint
     #     exit 1
     # fi
+
 }
 
 function stop_agent_docker() {
@@ -228,23 +230,23 @@ function stop_agent_docker() {
 
 
 echo "workpath: $WORKPATH"
-# echo "=================== Stop containers ===================="
-# stop_llm
+echo "=================== Stop containers ===================="
+stop_llm
 stop_agent_docker
 stop_dataprep
 
 cd $WORKPATH/tests
 
 # echo "=================== #1 Building docker images===================="
-# build_vllm_docker_image
-# build_dataprep_agent_images
+build_vllm_docker_image
+build_dataprep_agent_images
 
 #### for local test
 # build_agent_image_local
 # echo "=================== #1 Building docker images completed===================="
 
 # echo "=================== #2 Start vllm endpoint===================="
-# start_vllm_service_70B
+start_vllm_service_70B
 # echo "=================== #2 vllm endpoint started===================="
 
 # echo "=================== #3 Start dataprep and ingest data ===================="
@@ -257,12 +259,12 @@ start_agents
 validate_agent_service
 echo "=================== #4 Agent test passed ===================="
 
-# echo "=================== #5 Stop microservices ===================="
-# stop_agent_docker
-# stop_dataprep
-# stop_llm
-# echo "=================== #5 Microservices stopped===================="
+echo "=================== #5 Stop microservices ===================="
+stop_agent_docker
+stop_dataprep
+stop_llm
+echo "=================== #5 Microservices stopped===================="
 
-# echo y | docker system prune
+echo y | docker system prune
 
-# echo "ALL DONE!!"
+echo "ALL DONE!!"