Merge branch 'main' into deleteResource

opea-project · Aug 12, 2024 · cb981ce · cb981ce
2 parents e3561a7 + c0d2ba6
commit cb981ce
Show file tree

Hide file tree

Showing 30 changed files with 431 additions and 172 deletions.
diff --git a/.github/workflows/image-build-on-manual.yaml b/.github/workflows/image-build-on-manual.yaml
@@ -41,6 +41,7 @@ jobs:
           echo "nodes=$nodes" >> $GITHUB_OUTPUT
 
   image-build:
+    needs: get-build-matrix
     strategy:
       matrix:
         node: ${{ fromJSON(needs.get-build-matrix.outputs.nodes) }}

diff --git a/.github/workflows/scripts/e2e/gmc_gaudi_test.sh b/.github/workflows/scripts/e2e/gmc_gaudi_test.sh
@@ -11,6 +11,7 @@ USER_ID=$(whoami)
 LOG_PATH=/home/$(whoami)/logs
 CHATQNA_NAMESPACE="${APP_NAMESPACE}-chatqna"
 CHATQNA_DATAPREP_NAMESPACE="${APP_NAMESPACE}-chatqna-dataprep"
+CHATQNA_SWITCH_NAMESPACE="${APP_NAMESPACE}-chatqna-switch"
 CODEGEN_NAMESPACE="${APP_NAMESPACE}-codegen"
 CODETRANS_NAMESPACE="${APP_NAMESPACE}-codetrans"
 DOCSUM_NAMESPACE="${APP_NAMESPACE}-docsum"
@@ -22,21 +23,25 @@ function validate_gmc() {
     echo "validate chat-qna with dataprep"
     validate_chatqna_with_dataprep
 
-    echo "validate codegen"
-    validate_codegen
+    echo "validate chat-qna in switch mode"
+    validate_chatqna_in_switch
 
-    echo "validate codetrans"
-    validate_codetrans
+    # echo "validate codegen"
+    # validate_codegen
 
-    echo "validate docsum"
-    validate_docsum
+    # echo "validate codetrans"
+    # validate_codetrans
+
+    # echo "validate docsum"
+    # validate_docsum
 
     get_gmc_controller_logs
 }
 
 function cleanup_apps() {
     echo "clean up microservice-connector"
-    namespaces=("$CHATQNA_NAMESPACE" "$CHATQNA_DATAPREP_NAMESPACE" "$CODEGEN_NAMESPACE" "$CODETRANS_NAMESPACE" "$DOCSUM_NAMESPACE")
+    # namespaces=("$CHATQNA_NAMESPACE" "$CHATQNA_DATAPREP_NAMESPACE" "$CHATQNA_SWITCH_NAMESPACE" "$CODEGEN_NAMESPACE" "$CODETRANS_NAMESPACE" "$DOCSUM_NAMESPACE")
+    namespaces=("$CHATQNA_NAMESPACE" "$CHATQNA_DATAPREP_NAMESPACE" "$CHATQNA_SWITCH_NAMESPACE")
     for ns in "${namespaces[@]}"; do
         if kubectl get namespace $ns > /dev/null 2>&1; then
             echo "Deleting namespace: $ns"
@@ -101,7 +106,7 @@ function validate_chatqna() {
 }
 
 function validate_chatqna_with_dataprep() {
-    kubectl create ns $CHATQNA_DATAPREP_NAMESPACE
+   kubectl create ns $CHATQNA_DATAPREP_NAMESPACE
    sed -i "s|namespace: chatqa|namespace: $CHATQNA_DATAPREP_NAMESPACE|g"  $(pwd)/config/samples/chatQnA_dataprep_gaudi.yaml
    # workaround for issue #268
    yq -i '(.spec.nodes.root.steps[] | select ( .name == "Tgi")).internalService.config.MODEL_ID = "bigscience/bloom-560m"' $(pwd)/config/samples/chatQnA_dataprep_gaudi.yaml
@@ -178,6 +183,87 @@ function validate_chatqna_with_dataprep() {
    fi
 }
 
+function validate_chatqna_in_switch() {
+   kubectl create ns $CHATQNA_SWITCH_NAMESPACE
+   sed -i "s|namespace: switch|namespace: $CHATQNA_SWITCH_NAMESPACE|g"  $(pwd)/config/samples/chatQnA_switch_gaudi.yaml
+   # workaround for issue #268
+   yq -i '(.spec.nodes.root.steps[] | select ( .name == "Tgi")).internalService.config.MODEL_ID = "bigscience/bloom-560m"' $(pwd)/config/samples/chatQnA_switch_gaudi.yaml
+   kubectl apply -f $(pwd)/config/samples/chatQnA_switch_gaudi.yaml
+
+   # Wait until the router service is ready
+   echo "Waiting for the chatqa router service to be ready..."
+   wait_until_pod_ready "chatqna router" $CHATQNA_SWITCH_NAMESPACE "router-service"
+   output=$(kubectl get pods -n $CHATQNA_SWITCH_NAMESPACE)
+   echo $output
+
+   # deploy client pod for testing
+   kubectl create deployment client-test -n $CHATQNA_SWITCH_NAMESPACE --image=python:3.8.13 -- sleep infinity
+
+   # Wait until all pods are ready
+   wait_until_all_pod_ready $CHATQNA_SWITCH_NAMESPACE 300s
+   if [ $? -ne 0 ]; then
+       echo "Error Some pods are not ready!"
+       exit 1
+   fi
+
+   # giving time to populating data
+   sleep 90
+
+   kubectl get pods -n $CHATQNA_SWITCH_NAMESPACE
+   # send request to chatqnA
+   export CLIENT_POD=$(kubectl get pod -n $CHATQNA_SWITCH_NAMESPACE -l app=client-test -o jsonpath={.items..metadata.name})
+   echo "$CLIENT_POD"
+   accessUrl=$(kubectl get gmc -n $CHATQNA_SWITCH_NAMESPACE -o jsonpath="{.items[?(@.metadata.name=='switch')].status.accessUrl}")
+
+   # test the chatqna with model condition: "model-id":"intel" and "embedding-model-id":"small"
+   kubectl exec "$CLIENT_POD" -n $CHATQNA_SWITCH_NAMESPACE -- curl $accessUrl  -X POST  -d '{"text":"What is the revenue of Nike in 2023?", "model-id":"intel", "embedding-model-id":"small", "parameters":{"max_new_tokens":17, "do_sample": true}}' -H 'Content-Type: application/json' > $LOG_PATH/curl_chatqna_switch_intel.log
+   exit_code=$?
+   if [ $exit_code -ne 0 ]; then
+       echo "chatqna failed, please check the logs in ${LOG_PATH}!"
+       exit 1
+   fi
+
+   echo "Checking response results, make sure the output is reasonable. "
+   local status=false
+   if [[ -f $LOG_PATH/curl_chatqna_switch_intel.log ]] && \
+   [[ $(grep -c "[DONE]" $LOG_PATH/curl_chatqna_switch_intel.log) != 0 ]]; then
+       status=true
+   fi
+   if [ $status == false ]; then
+       if [[ -f $LOG_PATH/curl_chatqna_switch_intel.log ]]; then
+           cat $LOG_PATH/curl_chatqna_switch_intel.log
+       fi
+       echo "Response check failed, please check the logs in artifacts!"
+       exit 1
+   else
+       echo "Response check succeed!"
+   fi
+
+   # test the chatqna with model condition: "model-id":"llama" and "embedding-model-id":"large"
+   kubectl exec "$CLIENT_POD" -n $CHATQNA_SWITCH_NAMESPACE -- curl $accessUrl  -X POST  -d '{"text":"What is the revenue of Nike in 2023?", "model-id":"llama", "embedding-model-id":"large", "parameters":{"max_new_tokens":17, "do_sample": true}}' -H 'Content-Type: application/json' > $LOG_PATH/curl_chatqna_switch_llama.log
+   exit_code=$?
+   if [ $exit_code -ne 0 ]; then
+       echo "chatqna failed, please check the logs in ${LOG_PATH}!"
+       exit 1
+   fi
+
+   echo "Checking response results, make sure the output is reasonable. "
+   local status=false
+   if [[ -f $LOG_PATH/curl_chatqna_switch_llama.log ]] && \
+   [[ $(grep -c "[DONE]" $LOG_PATH/curl_chatqna_switch_llama.log) != 0 ]]; then
+       status=true
+   fi
+   if [ $status == false ]; then
+       if [[ -f $LOG_PATH/curl_chatqna_switch_llama.log ]]; then
+           cat $LOG_PATH/curl_chatqna_switch_llama.log
+       fi
+       echo "Response check failed, please check the logs in artifacts!"
+       exit 1
+   else
+       echo "Response check succeed!"
+   fi
+}
+
 function validate_codegen() {
    kubectl create ns $CODEGEN_NAMESPACE
    sed -i "s|namespace: codegen|namespace: $CODEGEN_NAMESPACE|g"  $(pwd)/config/samples/codegen_gaudi.yaml

diff --git a/.github/workflows/scripts/e2e/gmc_install.sh b/.github/workflows/scripts/e2e/gmc_install.sh
@@ -64,7 +64,7 @@ function cleanup_gmc() {
     if kubectl get namespace $SYSTEM_NAMESPACE > /dev/null 2>&1; then
         echo "Deleting namespace: $SYSTEM_NAMESPACE"
         kubectl delete namespace "$SYSTEM_NAMESPACE"
-        kubectl delete crd gmconnectors.gmc.opea.io
+        kubectl delete crd gmconnectors.gmc.opea.io || true
     else
         echo "Namespace $SYSTEM_NAMESPACE does not exist"
     fi

diff --git a/.github/workflows/scripts/e2e/gmc_xeon_test.sh b/.github/workflows/scripts/e2e/gmc_xeon_test.sh
@@ -11,6 +11,7 @@ USER_ID=$(whoami)
 LOG_PATH=/home/$(whoami)/logs
 CHATQNA_NAMESPACE="${APP_NAMESPACE}-chatqna"
 CHATQNA_DATAPREP_NAMESPACE="${APP_NAMESPACE}-chatqna-dataprep"
+CHATQNA_SWITCH_NAMESPACE="${APP_NAMESPACE}-chatqna-switch"
 CODEGEN_NAMESPACE="${APP_NAMESPACE}-codegen"
 CODETRANS_NAMESPACE="${APP_NAMESPACE}-codetrans"
 DOCSUM_NAMESPACE="${APP_NAMESPACE}-docsum"
@@ -22,21 +23,25 @@ function validate_gmc() {
     echo "validate chat-qna with dataprep"
     validate_chatqna_with_dataprep
 
-    echo "validate codegen"
-    validate_codegen
+    echo "validate chat-qna in switch mode"
+    validate_chatqna_in_switch
 
-    echo "validate codetrans"
-    validate_codetrans
+    # echo "validate codegen"
+    # validate_codegen
 
-    echo "validate docsum"
-    validate_docsum
+    # echo "validate codetrans"
+    # validate_codetrans
+
+    # echo "validate docsum"
+    # validate_docsum
 
     get_gmc_controller_logs
 }
 
 function cleanup_apps() {
     echo "clean up microservice-connector"
-    namespaces=("$CHATQNA_NAMESPACE" "$CHATQNA_DATAPREP_NAMESPACE" "$CODEGEN_NAMESPACE" "$CODETRANS_NAMESPACE" "$DOCSUM_NAMESPACE")
+    # namespaces=("$CHATQNA_NAMESPACE" "$CHATQNA_DATAPREP_NAMESPACE" "$CHATQNA_SWITCH_NAMESPACE" "$CODEGEN_NAMESPACE" "$CODETRANS_NAMESPACE" "$DOCSUM_NAMESPACE")
+    namespaces=("$CHATQNA_NAMESPACE" "$CHATQNA_DATAPREP_NAMESPACE" "$CHATQNA_SWITCH_NAMESPACE")
     for ns in "${namespaces[@]}"; do
         if kubectl get namespace $ns > /dev/null 2>&1; then
             echo "Deleting namespace: $ns"
@@ -103,7 +108,7 @@ function validate_chatqna() {
 }
 
 function validate_chatqna_with_dataprep() {
-    kubectl create ns $CHATQNA_DATAPREP_NAMESPACE
+   kubectl create ns $CHATQNA_DATAPREP_NAMESPACE
    sed -i "s|namespace: chatqa|namespace: $CHATQNA_DATAPREP_NAMESPACE|g"  $(pwd)/config/samples/chatQnA_dataprep_xeon.yaml
    # workaround for issue #268
    yq -i '(.spec.nodes.root.steps[] | select ( .name == "Tgi")).internalService.config.MODEL_ID = "bigscience/bloom-560m"' $(pwd)/config/samples/chatQnA_dataprep_xeon.yaml
@@ -180,6 +185,87 @@ function validate_chatqna_with_dataprep() {
    fi
 }
 
+function validate_chatqna_in_switch() {
+   kubectl create ns $CHATQNA_SWITCH_NAMESPACE
+   sed -i "s|namespace: switch|namespace: $CHATQNA_SWITCH_NAMESPACE|g"  $(pwd)/config/samples/chatQnA_switch_xeon.yaml
+   # workaround for issue #268
+   yq -i '(.spec.nodes.root.steps[] | select ( .name == "Tgi")).internalService.config.MODEL_ID = "bigscience/bloom-560m"' $(pwd)/config/samples/chatQnA_switch_xeon.yaml
+   kubectl apply -f $(pwd)/config/samples/chatQnA_switch_xeon.yaml
+
+   # Wait until the router service is ready
+   echo "Waiting for the chatqa router service to be ready..."
+   wait_until_pod_ready "chatqna router" $CHATQNA_SWITCH_NAMESPACE "router-service"
+   output=$(kubectl get pods -n $CHATQNA_SWITCH_NAMESPACE)
+   echo $output
+
+   # deploy client pod for testing
+   kubectl create deployment client-test -n $CHATQNA_SWITCH_NAMESPACE --image=python:3.8.13 -- sleep infinity
+
+   # Wait until all pods are ready
+   wait_until_all_pod_ready $CHATQNA_SWITCH_NAMESPACE 300s
+   if [ $? -ne 0 ]; then
+       echo "Error Some pods are not ready!"
+       exit 1
+   fi
+
+   # giving time to populating data
+   sleep 90
+
+   kubectl get pods -n $CHATQNA_SWITCH_NAMESPACE
+   # send request to chatqnA
+   export CLIENT_POD=$(kubectl get pod -n $CHATQNA_SWITCH_NAMESPACE -l app=client-test -o jsonpath={.items..metadata.name})
+   echo "$CLIENT_POD"
+   accessUrl=$(kubectl get gmc -n $CHATQNA_SWITCH_NAMESPACE -o jsonpath="{.items[?(@.metadata.name=='switch')].status.accessUrl}")
+
+   # test the chatqna with model condition: "model-id":"intel" and "embedding-model-id":"small"
+   kubectl exec "$CLIENT_POD" -n $CHATQNA_SWITCH_NAMESPACE -- curl $accessUrl  -X POST  -d '{"text":"What is the revenue of Nike in 2023?", "model-id":"intel", "embedding-model-id":"small", "parameters":{"max_new_tokens":17, "do_sample": true}}' -H 'Content-Type: application/json' > $LOG_PATH/curl_chatqna_switch_intel.log
+   exit_code=$?
+   if [ $exit_code -ne 0 ]; then
+       echo "chatqna failed, please check the logs in ${LOG_PATH}!"
+       exit 1
+   fi
+
+   echo "Checking response results, make sure the output is reasonable. "
+   local status=false
+   if [[ -f $LOG_PATH/curl_chatqna_switch_intel.log ]] && \
+   [[ $(grep -c "[DONE]" $LOG_PATH/curl_chatqna_switch_intel.log) != 0 ]]; then
+       status=true
+   fi
+   if [ $status == false ]; then
+       if [[ -f $LOG_PATH/curl_chatqna_switch_intel.log ]]; then
+           cat $LOG_PATH/curl_chatqna_switch_intel.log
+       fi
+       echo "Response check failed, please check the logs in artifacts!"
+       exit 1
+   else
+       echo "Response check succeed!"
+   fi
+
+   # test the chatqna with model condition: "model-id":"llama" and "embedding-model-id":"large"
+   kubectl exec "$CLIENT_POD" -n $CHATQNA_SWITCH_NAMESPACE -- curl $accessUrl  -X POST  -d '{"text":"What is the revenue of Nike in 2023?", "model-id":"llama", "embedding-model-id":"large", "parameters":{"max_new_tokens":17, "do_sample": true}}' -H 'Content-Type: application/json' > $LOG_PATH/curl_chatqna_switch_llama.log
+   exit_code=$?
+   if [ $exit_code -ne 0 ]; then
+       echo "chatqna failed, please check the logs in ${LOG_PATH}!"
+       exit 1
+   fi
+
+   echo "Checking response results, make sure the output is reasonable. "
+   local status=false
+   if [[ -f $LOG_PATH/curl_chatqna_switch_llama.log ]] && \
+   [[ $(grep -c "[DONE]" $LOG_PATH/curl_chatqna_switch_llama.log) != 0 ]]; then
+       status=true
+   fi
+   if [ $status == false ]; then
+       if [[ -f $LOG_PATH/curl_chatqna_switch_llama.log ]]; then
+           cat $LOG_PATH/curl_chatqna_switch_llama.log
+       fi
+       echo "Response check failed, please check the logs in artifacts!"
+       exit 1
+   else
+       echo "Response check succeed!"
+   fi
+}
+
 function validate_codegen() {
    kubectl create ns $CODEGEN_NAMESPACE
    sed -i "s|namespace: codegen|namespace: $CODEGEN_NAMESPACE|g"  $(pwd)/config/samples/codegen_xeon.yaml

diff --git a/helm-charts/chatqna/gaudi-values.yaml b/helm-charts/chatqna/gaudi-values.yaml
@@ -8,6 +8,8 @@ tei:
   resources:
     limits:
       habana.ai/gaudi: 1
+  securityContext:
+    readOnlyRootFilesystem: false
 
 # To override values in subchart tgi
 tgi:
@@ -17,8 +19,5 @@ tgi:
   resources:
     limits:
       habana.ai/gaudi: 1
-  extraArgs:
-    - "--max-input-length"
-    - "1024"
-    - "--max-total-tokens"
-    - "2048"
+  MAX_INPUT_LENGTH: "1024"
+  MAX_TOTAL_TOKENS: "2048"
diff --git a/helm-charts/codegen/gaudi-values.yaml b/helm-charts/codegen/gaudi-values.yaml
@@ -8,8 +8,5 @@ tgi:
   resources:
     limits:
       habana.ai/gaudi: 1
-  extraArgs:
-    - "--max-input-length"
-    - "1024"
-    - "--max-total-tokens"
-    - "2048"
+  MAX_INPUT_LENGTH: "1024"
+  MAX_TOTAL_TOKENS: "2048"
diff --git a/helm-charts/codetrans/gaudi-values.yaml b/helm-charts/codetrans/gaudi-values.yaml
@@ -8,8 +8,5 @@ tgi:
   resources:
     limits:
       habana.ai/gaudi: 1
-  extraArgs:
-    - "--max-input-length"
-    - "1024"
-    - "--max-total-tokens"
-    - "2048"
+  MAX_INPUT_LENGTH: "1024"
+  MAX_TOTAL_TOKENS: "2048"
diff --git a/helm-charts/common/data-prep/templates/deployment.yaml b/helm-charts/common/data-prep/templates/deployment.yaml
@@ -45,12 +45,6 @@ spec:
             - name: data-prep
               containerPort: 6007
               protocol: TCP
-            # The following need to be modified after GenAIComps bug #282 is resolved.
-            # https://github.com/opea-project/GenAIComps/issues/282
-            - containerPort: 6008
-              protocol: TCP
-            - containerPort: 6009
-              protocol: TCP
           volumeMounts:
             - mountPath: /tmp
               name: tmp

diff --git a/helm-charts/common/data-prep/templates/service.yaml b/helm-charts/common/data-prep/templates/service.yaml
@@ -10,11 +10,9 @@ metadata:
 spec:
   type: {{ .Values.service.type }}
   ports:
-    {{- range .Values.service.ports }}
-    - port: {{ .port }}
-      targetPort: {{ .targetPort }}
+    - port: {{ .Values.service.port }}
+      targetPort: 6007
       protocol: TCP
-      name: {{ .name }}
-    {{- end }}
+      name: data-prep
   selector:
     {{- include "data-prep.selectorLabels" . | nindent 4 }}
diff --git a/helm-charts/common/data-prep/templates/tests/test-pod.yaml b/helm-charts/common/data-prep/templates/tests/test-pod.yaml
@@ -18,12 +18,9 @@ spec:
       args:
         - |
           echo "test file" > /tmp/file1.txt;
-          {{- with index .Values.service.ports 0 }}
-          export port={{.port}};
-          {{- end }}
           max_retry=20;
           for ((i=1; i<=max_retry; i++)); do
-            curl http://{{ include "data-prep.fullname" . }}:$port/v1/dataprep -sS --fail-with-body \
+            curl http://{{ include "data-prep.fullname" . }}:{{ .Values.service.port }}/v1/dataprep -sS --fail-with-body \
             -X POST \
             -H "Content-Type: multipart/form-data" \
             -F "files=@/tmp/file1.txt" && break;

diff --git a/helm-charts/common/data-prep/values.yaml b/helm-charts/common/data-prep/values.yaml
@@ -38,19 +38,7 @@ securityContext:
 
 service:
   type: ClusterIP
-  # The following need to be modified after GenAIComps bug #282 is resolved.
-  # https://github.com/opea-project/GenAIComps/issues/282
-  ports:
-  # The default port for data prep service is 6007
-  - port: 6007
-    targetPort: 6007
-    name: data-prep
-  - port: 6008
-    targetPort: 6008
-    name: data-prep-get
-  - port: 6009
-    targetPort: 6009
-    name: data-prep-delete
+  port: 6007
 
 resources: {}
   # We usually recommend not to specify default resources and to leave this as a conscious