agent-starter-pack 0.3.3__py3-none-any.whl → 0.3.5__py3-none-any.whl

agents/crewai_coding_crew/notebooks/evaluating_langgraph_agent.ipynb

@@ -143,12 +143,8 @@
       },
       "outputs": [],
       "source": [
-        "%pip install --upgrade --user --quiet \"google-cloud-aiplatform[evaluation]\" \\\n",
-        "    \"langchain_google_vertexai\" \\\n",
-        "    \"langgraph\" \\\n",
-        "    \"cloudpickle==3.0.0\" \\\n",
-        "    \"pydantic==2.7.4\" \\\n",
-        "    \"requests\""
+        "%pip install \"langchain_google_vertexai\" \"langgraph\"\n",
+        "%pip install --upgrade --user --quiet \"google-cloud-aiplatform[evaluation]\""
       ]
     },
     {
@@ -273,6 +269,8 @@
       },
       "outputs": [],
       "source": [
+        "import json\n",
+        "\n",
         "# General\n",
         "import random\n",
         "import string\n",
@@ -355,7 +353,7 @@
         "                        }\n",
         "                    )\n",
         "\n",
-        "    final_output[\"predicted_trajectory\"] = function_calls\n",
+        "    final_output[\"predicted_trajectory\"] = json.dumps(function_calls)\n",
         "    return final_output\n",
         "\n",
         "\n",
@@ -365,6 +363,7 @@
         "    markdown += f\"{output['response']}\\n\\n\"\n",
         "\n",
         "    if output[\"predicted_trajectory\"]:\n",
+        "        output[\"predicted_trajectory\"] = json.loads(output[\"predicted_trajectory\"])\n",
         "        markdown += \"### Function Calls\\n\"\n",
         "        for call in output[\"predicted_trajectory\"]:\n",
         "            markdown += f\"- **Function**: `{call['tool_name']}`\\n\"\n",
@@ -381,7 +380,7 @@
         "    display(Markdown(\"### Summary Metrics\"))\n",
         "    display(metrics_df)\n",
         "\n",
-        "    display(Markdown(f\"### Row-wise Metrics\"))\n",
+        "    display(Markdown(\"### Row-wise Metrics\"))\n",
         "    display(eval_result.metrics_table)\n",
         "\n",
         "\n",
@@ -599,7 +598,7 @@
       "source": [
         "def router(\n",
         "    state: list[BaseMessage],\n",
-        ") -> Literal[\"get_product_details\", \"get_product_price\", \"__end__\"]:\n",
+        ") -> Literal[\"get_product_details\", \"get_product_price\", END]:\n",
         "    \"\"\"Initiates product details or price retrieval if the user asks for a product.\"\"\"\n",
         "    # Get the tool_calls from the last message in the conversation history.\n",
         "    tool_calls = state[-1].tool_calls\n",
@@ -614,7 +613,7 @@
         "            return \"get_product_details\"\n",
         "    else:\n",
         "        # End the conversation flow.\n",
-        "        return \"__end__\""
+        "        return END"
       ]
     },
     {
@@ -636,7 +635,7 @@
       },
       "outputs": [],
       "source": [
-        "llm = \"gemini-1.5-pro\""
+        "llm = \"gemini-2.0-flash\""
       ]
     },
     {
@@ -899,6 +898,7 @@
         "    dataset=eval_sample_dataset,\n",
         "    metrics=single_tool_usage_metrics,\n",
         "    experiment=EXPERIMENT_NAME,\n",
+        "    output_uri_prefix=BUCKET_URI + \"/single-metric-eval\",\n",
         ")\n",
         "\n",
         "single_tool_call_eval_result = single_tool_call_eval_task.evaluate(\n",
@@ -1003,7 +1003,10 @@
         "EXPERIMENT_RUN = f\"trajectory-{get_id()}\"\n",
         "\n",
         "trajectory_eval_task = EvalTask(\n",
-        "    dataset=eval_sample_dataset, metrics=trajectory_metrics, experiment=EXPERIMENT_NAME\n",
+        "    dataset=eval_sample_dataset,\n",
+        "    metrics=trajectory_metrics,\n",
+        "    experiment=EXPERIMENT_NAME,\n",
+        "    output_uri_prefix=BUCKET_URI + \"/multiple-metric-eval\",\n",
         ")\n",
         "\n",
         "trajectory_eval_result = trajectory_eval_task.evaluate(\n",
@@ -1107,7 +1110,10 @@
         "EXPERIMENT_RUN = f\"response-{get_id()}\"\n",
         "\n",
         "response_eval_task = EvalTask(\n",
-        "    dataset=eval_sample_dataset, metrics=response_metrics, experiment=EXPERIMENT_NAME\n",
+        "    dataset=eval_sample_dataset,\n",
+        "    metrics=response_metrics,\n",
+        "    experiment=EXPERIMENT_NAME,\n",
+        "    output_uri_prefix=BUCKET_URI + \"/response-metric-eval\",\n",
         ")\n",
         "\n",
         "response_eval_result = response_eval_task.evaluate(\n",
@@ -1292,6 +1298,7 @@
         "    dataset=eval_sample_dataset,\n",
         "    metrics=response_tool_metrics,\n",
         "    experiment=EXPERIMENT_NAME,\n",
+        "    output_uri_prefix=BUCKET_URI + \"/reasoning-metric-eval\",\n",
         ")\n",
         "\n",
         "response_eval_tool_result = response_eval_tool_task.evaluate(\n",
@@ -1436,15 +1443,21 @@
         "        ],\n",
         "    ],\n",
         "    \"response\": [\n",
-        "        500,\n",
-        "        50,\n",
+        "        \"500\",\n",
+        "        \"50\",\n",
         "        \"A super fast and light usb charger\",\n",
-        "        100,\n",
+        "        \"100\",\n",
         "        \"A voice-controlled smart speaker that plays music, sets alarms, and controls smart home devices.\",\n",
         "    ],\n",
         "}\n",
         "\n",
-        "byod_eval_sample_dataset = pd.DataFrame(byod_eval_data)"
+        "byod_eval_sample_dataset = pd.DataFrame(byod_eval_data)\n",
+        "byod_eval_sample_dataset[\"predicted_trajectory\"] = byod_eval_sample_dataset[\n",
+        "    \"predicted_trajectory\"\n",
+        "].apply(json.dumps)\n",
+        "byod_eval_sample_dataset[\"reference_trajectory\"] = byod_eval_sample_dataset[\n",
+        "    \"reference_trajectory\"\n",
+        "].apply(json.dumps)"
       ]
     },
     {

agents/langgraph_base_react/notebooks/evaluating_langgraph_agent.ipynb

@@ -143,12 +143,8 @@
       },
       "outputs": [],
       "source": [
-        "%pip install --upgrade --user --quiet \"google-cloud-aiplatform[evaluation]\" \\\n",
-        "    \"langchain_google_vertexai\" \\\n",
-        "    \"langgraph\" \\\n",
-        "    \"cloudpickle==3.0.0\" \\\n",
-        "    \"pydantic==2.7.4\" \\\n",
-        "    \"requests\""
+        "%pip install \"langchain_google_vertexai\" \"langgraph\"\n",
+        "%pip install --upgrade --user --quiet \"google-cloud-aiplatform[evaluation]\""
       ]
     },
     {
@@ -273,6 +269,8 @@
       },
       "outputs": [],
       "source": [
+        "import json\n",
+        "\n",
         "# General\n",
         "import random\n",
         "import string\n",
@@ -355,7 +353,7 @@
         "                        }\n",
         "                    )\n",
         "\n",
-        "    final_output[\"predicted_trajectory\"] = function_calls\n",
+        "    final_output[\"predicted_trajectory\"] = json.dumps(function_calls)\n",
         "    return final_output\n",
         "\n",
         "\n",
@@ -365,6 +363,7 @@
         "    markdown += f\"{output['response']}\\n\\n\"\n",
         "\n",
         "    if output[\"predicted_trajectory\"]:\n",
+        "        output[\"predicted_trajectory\"] = json.loads(output[\"predicted_trajectory\"])\n",
         "        markdown += \"### Function Calls\\n\"\n",
         "        for call in output[\"predicted_trajectory\"]:\n",
         "            markdown += f\"- **Function**: `{call['tool_name']}`\\n\"\n",
@@ -381,7 +380,7 @@
         "    display(Markdown(\"### Summary Metrics\"))\n",
         "    display(metrics_df)\n",
         "\n",
-        "    display(Markdown(f\"### Row-wise Metrics\"))\n",
+        "    display(Markdown(\"### Row-wise Metrics\"))\n",
         "    display(eval_result.metrics_table)\n",
         "\n",
         "\n",
@@ -599,7 +598,7 @@
       "source": [
         "def router(\n",
         "    state: list[BaseMessage],\n",
-        ") -> Literal[\"get_product_details\", \"get_product_price\", \"__end__\"]:\n",
+        ") -> Literal[\"get_product_details\", \"get_product_price\", END]:\n",
         "    \"\"\"Initiates product details or price retrieval if the user asks for a product.\"\"\"\n",
         "    # Get the tool_calls from the last message in the conversation history.\n",
         "    tool_calls = state[-1].tool_calls\n",
@@ -614,7 +613,7 @@
         "            return \"get_product_details\"\n",
         "    else:\n",
         "        # End the conversation flow.\n",
-        "        return \"__end__\""
+        "        return END"
       ]
     },
     {
@@ -636,7 +635,7 @@
       },
       "outputs": [],
       "source": [
-        "llm = \"gemini-1.5-pro\""
+        "llm = \"gemini-2.0-flash\""
       ]
     },
     {
@@ -899,6 +898,7 @@
         "    dataset=eval_sample_dataset,\n",
         "    metrics=single_tool_usage_metrics,\n",
         "    experiment=EXPERIMENT_NAME,\n",
+        "    output_uri_prefix=BUCKET_URI + \"/single-metric-eval\",\n",
         ")\n",
         "\n",
         "single_tool_call_eval_result = single_tool_call_eval_task.evaluate(\n",
@@ -1003,7 +1003,10 @@
         "EXPERIMENT_RUN = f\"trajectory-{get_id()}\"\n",
         "\n",
         "trajectory_eval_task = EvalTask(\n",
-        "    dataset=eval_sample_dataset, metrics=trajectory_metrics, experiment=EXPERIMENT_NAME\n",
+        "    dataset=eval_sample_dataset,\n",
+        "    metrics=trajectory_metrics,\n",
+        "    experiment=EXPERIMENT_NAME,\n",
+        "    output_uri_prefix=BUCKET_URI + \"/multiple-metric-eval\",\n",
         ")\n",
         "\n",
         "trajectory_eval_result = trajectory_eval_task.evaluate(\n",
@@ -1107,7 +1110,10 @@
         "EXPERIMENT_RUN = f\"response-{get_id()}\"\n",
         "\n",
         "response_eval_task = EvalTask(\n",
-        "    dataset=eval_sample_dataset, metrics=response_metrics, experiment=EXPERIMENT_NAME\n",
+        "    dataset=eval_sample_dataset,\n",
+        "    metrics=response_metrics,\n",
+        "    experiment=EXPERIMENT_NAME,\n",
+        "    output_uri_prefix=BUCKET_URI + \"/response-metric-eval\",\n",
         ")\n",
         "\n",
         "response_eval_result = response_eval_task.evaluate(\n",
@@ -1292,6 +1298,7 @@
         "    dataset=eval_sample_dataset,\n",
         "    metrics=response_tool_metrics,\n",
         "    experiment=EXPERIMENT_NAME,\n",
+        "    output_uri_prefix=BUCKET_URI + \"/reasoning-metric-eval\",\n",
         ")\n",
         "\n",
         "response_eval_tool_result = response_eval_tool_task.evaluate(\n",
@@ -1436,15 +1443,21 @@
         "        ],\n",
         "    ],\n",
         "    \"response\": [\n",
-        "        500,\n",
-        "        50,\n",
+        "        \"500\",\n",
+        "        \"50\",\n",
         "        \"A super fast and light usb charger\",\n",
-        "        100,\n",
+        "        \"100\",\n",
         "        \"A voice-controlled smart speaker that plays music, sets alarms, and controls smart home devices.\",\n",
         "    ],\n",
         "}\n",
         "\n",
-        "byod_eval_sample_dataset = pd.DataFrame(byod_eval_data)"
+        "byod_eval_sample_dataset = pd.DataFrame(byod_eval_data)\n",
+        "byod_eval_sample_dataset[\"predicted_trajectory\"] = byod_eval_sample_dataset[\n",
+        "    \"predicted_trajectory\"\n",
+        "].apply(json.dumps)\n",
+        "byod_eval_sample_dataset[\"reference_trajectory\"] = byod_eval_sample_dataset[\n",
+        "    \"reference_trajectory\"\n",
+        "].apply(json.dumps)"
       ]
     },
     {

src/base_template/Makefile

@@ -7,12 +7,12 @@ test:
 
 playground:
 {%- if "adk" in cookiecutter.tags %}
-	@echo "+-----------------------------------------------------------------------------+"
-	@echo "| Starting ADK Web Server via 'adk web' command.                              |"
-	@echo "|                                                                             |"
-	@echo "| Please select the app folder to start interacting with your agent.          |"
-	@echo "+-----------------------------------------------------------------------------+"
-	uv run adk web
+	@echo "\033[1;32m===============================================================================\033[0m"
+	@echo "\033[1;32m| Starting ADK Web Server via 'adk web' command.                              |\033[0m"
+	@echo "\033[1;32m|                                                                             |\033[0m"
+	@echo "\033[1;32m| IMPORTANT: Select the 'app' folder to interact with your agent.             |\033[0m"
+	@echo "\033[1;32m===============================================================================\033[0m"
+	uv run adk web --port 8501
 {%- else %}
 {%- if cookiecutter.deployment_target == 'cloud_run' %}
 	uv run uvicorn app.server:app --host 0.0.0.0 --port 8000 --reload &
@@ -26,29 +26,42 @@ playground:
 
 backend:
 {%- if cookiecutter.deployment_target == 'cloud_run' %}
-	uv run uvicorn app.server:app --host 0.0.0.0 --port 8000 --reload
+	PROJECT_ID=$$(gcloud config get-value project) && \
+	gcloud run deploy {{cookiecutter.project_name}} \
+		--source . \
+		--memory "4Gi" \
+		--project $$PROJECT_ID \
+		--region "us-central1" \
+		--no-allow-unauthenticated \
+		--labels "created-by=adk" \
+		--set-env-vars \
+		"COMMIT_SHA=$(shell git rev-parse HEAD){%- if cookiecutter.data_ingestion %}{%- if cookiecutter.datastore_type == "vertex_ai_search" %},DATA_STORE_ID={{cookiecutter.project_name}}-datastore,DATA_STORE_REGION=us{%- elif cookiecutter.datastore_type == "vertex_ai_vector_search" %},VECTOR_SEARCH_INDEX={{cookiecutter.project_name}}-vector-search,VECTOR_SEARCH_INDEX_ENDPOINT={{cookiecutter.project_name}}-vector-search-endpoint,VECTOR_SEARCH_BUCKET=$$PROJECT_ID-{{cookiecutter.project_name}}-vs{%- endif %}{%- endif %}"
 {%- elif cookiecutter.deployment_target == 'agent_engine' %}
-	# Export dependencies to requirements file using uv export (preferred method), otherwise fall back to uv pip freeze
+	# Export dependencies to requirements file using uv export.
 	uv export --no-hashes --no-sources --no-header --no-dev --no-emit-project --no-annotate --frozen > .requirements.txt 2>/dev/null || \
-	uv pip freeze --exclude-editable > .requirements.txt && uv run app/agent_engine_app.py
+	uv export --no-hashes --no-sources --no-header --no-dev --no-emit-project --frozen > .requirements.txt && uv run app/agent_engine_app.py
 {%- endif %}
+{%- if cookiecutter.deployment_target == 'cloud_run' %}
 
-{% if cookiecutter.deployment_target == 'cloud_run' -%}
-ui:
+local-backend:
+	uv run uvicorn app.server:app --host 0.0.0.0 --port 8000 --reload
+{%- endif %}
+{%- if cookiecutter.deployment_target == 'cloud_run' %}
 {%- if cookiecutter.agent_name == 'live_api' %}
+
+ui:
 	npm --prefix frontend start
-{%- else %}
-	uv run streamlit run streamlit/streamlit_app.py --browser.serverAddress=localhost --server.enableCORS=false --server.enableXsrfProtection=false
 {%- endif %}
 {%- endif %}
 
 setup-dev-env:
-	@if [ -z "$$PROJECT_ID" ]; then echo "Error: PROJECT_ID environment variable is not set"; exit 1; fi
+	PROJECT_ID=$$(gcloud config get-value project) && \
 	(cd deployment/terraform/dev && terraform init && terraform apply --var-file vars/env.tfvars --var dev_project_id=$$PROJECT_ID --auto-approve)
-{% if cookiecutter.data_ingestion %}
+
+{%- if cookiecutter.data_ingestion %}
+
 data-ingestion:
-	@if [ -z "$$PROJECT_ID" ]; then echo "Error: PROJECT_ID environment variable is not set"; exit 1; fi
-	$(MAKE) install
+	PROJECT_ID=$$(gcloud config get-value project) && \
 	(cd data_ingestion && uv run data_ingestion_pipeline/submit_pipeline.py \
 		--project-id=$$PROJECT_ID \
 		--region="us-central1" \

src/base_template/README.md

@@ -48,11 +48,16 @@ make install && make playground
 | `make install`       | Install all required dependencies using uv                                                  |
 {%- if cookiecutter.deployment_target == 'cloud_run' %}
 | `make playground`    | Launch local development environment with backend and frontend{%- if "adk" in cookiecutter.tags %} - leveraging `adk web` command. {%- endif %}|
-| `make backend`       | Start backend server only |
-| `make ui`            | Launch Streamlit frontend without local backend |
+| `make backend`       | Deploy agent to Cloud Run |
+| `make local-backend`       | Deploy agent to Cloud Run |
+{%- if cookiecutter.deployment_target == 'cloud_run' %}
+{%- if cookiecutter.agent_name == 'live_api' %}
+| `make ui`       | Launch Agent Playground front-end only |
+{%- endif %}
+{%- endif %}
 {%- elif cookiecutter.deployment_target == 'agent_engine' %}
 | `make playground`    | Launch Streamlit interface for testing agent locally and remotely |
-| `make backend`       | Deploy agent to Agent Engine service |
+| `make backend`       | Deploy agent to Agent Engine |
 {%- endif %}
 | `make test`          | Run unit and integration tests                                                              |
 | `make lint`          | Run code quality checks (codespell, ruff, mypy)                                             |
@@ -158,24 +163,12 @@ This template follows a "bring your own agent" approach - you focus on your busi
 
 ### Dev Environment
 
-{%- if cookiecutter.deployment_target == 'agent_engine' %}
 You can test deployment towards a Dev Environment using the following command:
 
 ```bash
 gcloud config set project <your-dev-project-id>
 make backend
 ```
-{%- elif cookiecutter.deployment_target == 'cloud_run' %}
-Deploy the application directly to Cloud Run from your source code using the following `gcloud` command:
-
-```bash
-gcloud run deploy genai-app-sample \
-  --source . \
-  --project YOUR_PROJECT_ID \
-  --region YOUR_GCP_REGION \
-  --memory "4Gi" \
-```
-Replace `YOUR_PROJECT_ID` with your Google Cloud project ID and `YOUR_GCP_REGION` with the desired region (e.g., `us-central1`). Adjust memory and other flags as needed for your environment.
 {% if cookiecutter.agent_name == 'live_api' %}
 **Accessing the Deployed Backend Locally:**
 
@@ -190,7 +183,6 @@ To connect your local frontend (`make ui`) to the backend deployed on Cloud Run,
 
 2.  **Connect Frontend:** Your deployed backend is now accessible locally at `http://localhost:8000`. Point your Streamlit UI to this address.
 {%- endif %}
-{%- endif %}
 
 The repository includes a Terraform configuration for the setup of the Dev Google Cloud project.
 See [deployment/README.md](deployment/README.md) for instructions.
@@ -201,8 +193,8 @@ The repository includes a Terraform configuration for the setup of a production
 
 {% if cookiecutter.agent_name != 'live_api' %}
 ## Monitoring and Observability
-
-> You can use [this Looker Studio dashboard](https://lookerstudio.google.com/c/reporting/fa742264-4b4b-4c56-81e6-a667dd0f853f/page/tEnnC) template for visualizing events being logged in BigQuery. See the "Setup Instructions" tab to getting started.
+> You can use [this Looker Studio dashboard]({%- if "adk" in cookiecutter.tags %}https://lookerstudio.google.com/reporting/46b35167-b38b-4e44-bd37-701ef4307418/page/tEnnC{%- else %}https://lookerstudio.google.com/c/reporting/fa742264-4b4b-4c56-81e6-a667dd0f853f/page/tEnnC{%- endif %}
+) template for visualizing events being logged in BigQuery. See the "Setup Instructions" tab to getting started.
 
 The application uses OpenTelemetry for comprehensive observability with all events being sent to Google Cloud Trace and Logging for monitoring and to BigQuery for long term storage. 
 {%- endif %}

src/base_template/deployment/cd/deploy-to-prod.yaml

@@ -66,6 +66,9 @@ steps:
       - "40"
       - "--service-account"
       - "${_CLOUD_RUN_APP_SA_EMAIL}"
+      - "--session-affinity"
+      - "--labels"
+      - "created-by=adk"
       - "--set-env-vars"
       - "COMMIT_SHA=${COMMIT_SHA}{%- if cookiecutter.data_ingestion %}{%- if cookiecutter.datastore_type == "vertex_ai_search" %},DATA_STORE_ID=${_DATA_STORE_ID},DATA_STORE_REGION=${_DATA_STORE_REGION}{%- elif cookiecutter.datastore_type == "vertex_ai_vector_search" %},VECTOR_SEARCH_INDEX=${_VECTOR_SEARCH_INDEX},VECTOR_SEARCH_INDEX_ENDPOINT=${_VECTOR_SEARCH_INDEX_ENDPOINT},VECTOR_SEARCH_BUCKET=${_VECTOR_SEARCH_BUCKET}{%- endif %}{%- endif %}"
 {%- elif cookiecutter.deployment_target == 'agent_engine' %}

src/base_template/deployment/cd/staging.yaml

@@ -80,6 +80,9 @@ steps:
       - "40"
       - "--service-account"
       - "${_CLOUD_RUN_APP_SA_EMAIL}"
+      - "--session-affinity"
+      - "--labels"
+      - "created-by=adk"
       - "--set-env-vars"
       - "COMMIT_SHA=${COMMIT_SHA}{%- if cookiecutter.data_ingestion %}{%- if cookiecutter.datastore_type == "vertex_ai_search" %},DATA_STORE_ID=${_DATA_STORE_ID},DATA_STORE_REGION=${_DATA_STORE_REGION}{%- elif cookiecutter.datastore_type == "vertex_ai_vector_search" %},VECTOR_SEARCH_INDEX=${_VECTOR_SEARCH_INDEX},VECTOR_SEARCH_INDEX_ENDPOINT=${_VECTOR_SEARCH_INDEX_ENDPOINT},VECTOR_SEARCH_BUCKET=${_VECTOR_SEARCH_BUCKET}{%- endif %}{%- endif %}"
 

src/frontends/live_api_react/frontend/package-lock.json

@@ -10046,9 +10046,9 @@
       }
     },
     "node_modules/http-proxy-middleware": {
-      "version": "2.0.7",
-      "resolved": "https://registry.npmjs.org/http-proxy-middleware/-/http-proxy-middleware-2.0.7.tgz",
-      "integrity": "sha512-fgVY8AV7qU7z/MmXJ/rxwbrtQH4jBQ9m7kp3llF0liB7glmFeVZFBepQb32T3y8n8k2+AEYuMPCpinYW+/CuRA==",
+      "version": "2.0.9",
+      "resolved": "https://registry.npmjs.org/http-proxy-middleware/-/http-proxy-middleware-2.0.9.tgz",
+      "integrity": "sha512-c1IyJYLYppU574+YI7R4QyX2ystMtVXZwIdzazUIPIJsHuWNd+mho2j+bKoHftndicGj9yh+xjd+l0yj7VeT1Q==",
       "license": "MIT",
       "dependencies": {
         "@types/http-proxy": "^1.17.8",

src/resources/idx/idx-template.json

@@ -12,7 +12,7 @@
     },
     {
       "id": "google_cloud_project_id",
-      "name": "Your Goole Cloud Project ID",
+      "name": "Your Google Cloud Project ID",
       "description": "The Google Cloud Platform Project ID where resources will be managed.",
       "type": "string",
       "required": true