Updated prompts and tests

.github/workflows/cloud-test-router.yml

@@ -0,0 +1,146 @@
+name: Test Router Services with Cloud SQL
+
+on:
+  push:
+    branches: [ master ]
+  pull_request:
+    branches: [ master ]
+
+jobs:
+  test-router:
+    # 1. REQUIRED: Set permissions to get the GitHub OIDC token
+    permissions:
+      contents: 'read'
+      id-token: 'write' # Grants permission to fetch the OIDC token for WIF
+
+    runs-on: ubuntu-latest
+
+    # Global environment variables for the job
+    env:
+      # Application path setup
+      PYTHONPATH: ${{ github.workspace }}/backend:${{ github.workspace }}/backend/router:${{ github.workspace }}/backend/embeddings:${{ github.workspace }}/backend/database
+
+      # *** WIF & Cloud SQL IAM ENV VARS ***
+      SERVICE_ACCOUNT_EMAIL: ${{ secrets.GCP_SERVICE_ACCOUNT }}
+      INSTANCE_CONNECTION_NAME: ${{ secrets.INSTANCE_CONNECTION_NAME }}
+
+      # Construct the DATABASE_URL for IAM Auth: user is the SA email, password is empty, host is 127.0.0.1 (proxy)
+      DATABASE_URL: postgresql://${{ secrets.GCP_SERVICE_ACCOUNT }}:@127.0.0.1:5432/${{ secrets.DB_NAME }}
+
+    steps:
+    - name: Checkout code
+      uses: actions/checkout@v4
+
+    - name: Set up Python
+      uses: actions/setup-python@v4
+      with:
+        python-version: '3.11'
+
+    - name: Cache pip dependencies
+      uses: actions/cache@v3
+      with:
+        path: ~/.cache/pip
+        key: ${{ runner.os }}-pip-${{ hashFiles('backend/**/pyproject.toml', 'backend/**/requirements.txt') }}
+        restore-keys: |
+          ${{ runner.os }}-pip-
+
+
+
+    # 5. AUTHENTICATE TO GOOGLE CLOUD VIA WIF
+    - name: 'Authenticate to Google Cloud'
+      id: 'auth'
+      uses: 'google-github-actions/auth@v2'
+      with:
+        workload_identity_provider: ${{ secrets.GCP_WORKLOAD_IDENTITY_PROVIDER }}
+        service_account: ${{ secrets.GCP_SERVICE_ACCOUNT }}
+        # Credentials are now set as Application Default Credentials (ADC)
+
+    # 6. START CLOUD SQL PROXY USING IAM AUTH
+    - name: 'Start Cloud SQL Proxy'
+      run: |
+        wget https://storage.googleapis.com/cloud-sql-connectors/cloud-sql-proxy/v2.10.0/cloud-sql-proxy.linux.amd64 -O cloud-sql-proxy
+        chmod +x cloud-sql-proxy
+        # --auto-iam-authn uses the ADC established by the 'Authenticate' step
+        ./cloud-sql-proxy --auto-iam-authn ${{ env.INSTANCE_CONNECTION_NAME }} &
+        echo $! > cloud-sql-proxy.pid
+        sleep 5 # Wait for the proxy to initialize
+
+    - name: Create test environment file
+      working-directory: ./backend
+      run: |
+        cat > .env << EOF
+        RATING_INFERENCE_KEY=${{ secrets.RATING_INFERENCE_KEY }}
+        INFERENCE_URL=https://inference.geist.im
+        EMBEDDINGS_URL=https://embeddings.geist.im
+        HARMONY_ENABLED=false
+        LOG_LEVEL=INFO
+        # Use the WIF-constructed DATABASE_URL from the env block
+        DATABASE_URL=${{ env.DATABASE_URL }} 
+        EOF
+
+
+    - name: Install Python dependencies
+      working-directory: ./backend
+      run: |
+        python -m pip install --upgrade pip
+        pip install fastapi httpx uvicorn sse-starlette python-multipart python-dotenv
+        pip install pytest pytest-asyncio sentence_transformers sqlalchemy alembic python-dateutil psycopg2
+
+    - name: Start router service
+      working-directory: ./backend/router
+      run: |
+        python main.py > router.log 2>&1 &
+        echo $! > router.pid
+        sleep 10
+      env:
+        SKIP_TEST_SAVING: true
+        RATING_INFERENCE_KEY: ${{ secrets.RATING_INFERENCE_KEY }}
+        USE_REMOTE_INFERENCE: false
+        INFERENCE_URL: https://inference.geist.im
+        EMBEDDINGS_URL: https://embeddings.geist.im
+        LOG_LEVEL: INFO
+        # Note: App should read DATABASE_URL from .env file created above
+
+    - name: Debug router log
+      run: cat backend/router/router.log || true
+
+    - name: Wait for router to be ready
+      run: |
+        timeout 60 bash -c 'until curl -f http://localhost:8000/health; do sleep 2; done'
+
+    - name: Run streaming tests (test_conversation.py)
+      working-directory: ./backend/router
+      run: |
+        echo "=== Starting streaming tests ==="
+        python test_conversation.py
+        echo "=== Streaming tests completed ==="
+      # No need to explicitly pass DATABASE_URL here if the test also reads the .env file
+
+    - name: Run health check tests
+      working-directory: ./backend/router
+      run: |
+        python test_health_endpoint.py
+
+    - name: Save router logs
+      if: always()
+      working-directory: ./backend/router
+      run: |
+        echo "=== Router Logs ==="
+        cat router.log || echo "No router.log found"
+        echo "=== End Router Logs ==="
+
+    - name: Cleanup
+      if: always()
+      run: |
+        # Cleanup router service
+        if [ -f backend/router/router.pid ]; then
+          kill $(cat backend/router/router.pid) 2>/dev/null || true
+          rm -f backend/router/router.pid
+        fi
+        pkill -f "python main.py" || true
+
+        # Cleanup Cloud SQL Proxy
+        if [ -f cloud-sql-proxy.pid ]; then
+          kill $(cat cloud-sql-proxy.pid) 2>/dev/null || true
+          rm -f cloud-sql-proxy.pid
+        fi

backend/database/migrate.py

@@ -9,6 +9,7 @@
 import subprocess
 import logging
 from pathlib import Path
+from dotenv import load_dotenv
 
 # Add the backend directory to the Python path
 sys.path.append(os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
@@ -24,13 +25,27 @@
 
 def run_alembic_command(command: str, *args):
     """Run an alembic command with proper environment setup"""
+    try:
+        # Get the directory where this config.py file is located
+        env_file = Path(__file__).parent
+        # Go up one directory to find the .env file
+        env_file = env_file.parent / ".env"
+        print(f"Loading .env file from: {env_file}")
+        if env_file.exists():
+            load_dotenv(env_file)
+            print(f"Loaded environment variables from: {env_file}")
+        else:
+            print(f"No .env file found at: {env_file}")
+    except ImportError:
+        print("python-dotenv not installed, skipping .env file loading")
     try:
         # Set environment variables
         env = os.environ.copy()
         env['DATABASE_URL'] = os.getenv(
             'DATABASE_URL', 
             'postgresql://postgres:password@localhost:5433/test-storage'
         )
+        print(f"Using DATABASE_URL: {env['DATABASE_URL']}")
         # Change to the database directory
         db_dir = Path(__file__).parent
         os.chdir(db_dir)

backend/database/migrations/versions/0004 add columns to response.py

@@ -0,0 +1,27 @@
+"""Add first_token_time and num_tool_calls columns to conversation_response
+
+Revision ID: 0004
+Revises: 0003
+Create Date: 2024-06-09 00:00:00.000000
+
+"""
+
+from alembic import op
+import sqlalchemy as sa
+
+# revision identifiers, used by Alembic.
+revision = '0004'
+down_revision = '0003'
+branch_labels = None
+depends_on = None
+
+def upgrade() -> None:
+    # Add first_token_time (float) and num_tool_calls (integer) to conversation_response
+    op.add_column('conversation_response', sa.Column('first_token_time', sa.Float(), nullable=True))
+    op.add_column('conversation_response', sa.Column('num_tool_calls', sa.Integer(), nullable=True))
+
+def downgrade() -> None:
+    # Remove the two columns in downgrade
+    op.drop_column('conversation_response', 'num_tool_calls')
+    op.drop_column('conversation_response', 'first_token_time')
+

backend/database/migrations/versions/0005_add_test_run_time_to_conversation_.py

@@ -0,0 +1,29 @@
+"""Add test_run_time to conversation_response
+
+Revision ID: 6e6db6b65802
+Revises: 0004
+Create Date: 2025-10-27 10:31:49.902161
+
+"""
+from alembic import op
+import sqlalchemy as sa
+
+# revision identifiers, used by Alembic.
+revision = '6e6db6b65802'
+down_revision = '0004'
+branch_labels = None
+depends_on = None
+
+
+def upgrade() -> None:
+    # ### commands auto generated by Alembic - please adjust! ###
+    op.add_column('conversation_response', sa.Column('test_run_time', sa.DateTime(timezone=True), nullable=True))
+    # ### end Alembic commands ###
+
+
+def downgrade() -> None:
+    # ### commands auto generated by Alembic - please adjust! ###
+    op.drop_column('conversation_response', 'test_run_time')
+    # ### end Alembic commands ###
+
+

backend/database/models.py

@@ -36,6 +36,9 @@ class ConversationResponse(Base):
     rationality = Column(Float, nullable=True)  # Rationality score
     coherency = Column(Float, nullable=True)    # Coherency score
     elapsed_time = Column(Float, nullable=True)  # Response time in seconds
+    first_token_time = Column(Float, nullable=True)  # Time to first token
+    num_tool_calls = Column(Integer, nullable=True)  # Number of tool calls
+    test_run_time = Column(DateTime(timezone=True), nullable=True)  # Timestamp for test suite iteration
 
     # Foreign key to conversation (many responses belong to one conversation)
     conversation_id = Column(Integer, ForeignKey('conversation.internal_id', ondelete='CASCADE'), nullable=True)

backend/docker-compose.chris.yml

@@ -3,7 +3,7 @@ services:
     build: ./router
     ports:
       - "0.0.0.0:8000:8000"  # Bind to all interfaces
-      - "0.0.0.0:8443:8443"# HTTPS port (uncomment if using SSL)
+      - "0.0.0.0:8443:8443" # HTTPS port (uncomment if using SSL)
     environment:
       - LOG_LEVEL=DEBUG
       - HARMONY_REASONING_EFFORT=low
@@ -14,12 +14,13 @@ services:
       - PYTHONUNBUFFERED=1
       - PYTHONDONTWRITEBYTECODE=1
       - WATCHDOG_POLLING=true
-
-      - OPENAI_URL=https://api.openai.com
-      - USE_REMOTE_INFERENCE=${USE_REMOTE_INFERENCE}
-      - OPENAI_KEY=${OPENAI_KEY}
+      - RATING_INFERENCE_URL=${RATING_INFERENCE_URL}
+      - RATING_INFERENCE_KEY=${RATING_INFERENCE_KEY}
+      - RATING_INFERENCE_MODEL=${RATING_INFERENCE_MODEL}
       - REMOTE_INFERENCE_KEY=${REMOTE_INFERENCE_KEY}
-      - USE_REMOTE_INFERENCE=true
+      - USE_REMOTE_INFERENCE=${USE_REMOTE_INFERENCE}
+      - REMOTE_INFERENCE_MODEL=${REMOTE_INFERENCE_MODEL}
+      - REMOTE_INFERENCE_URL=${REMOTE_INFERENCE_URL}
       - BRAVE_API_KEY=${BRAVE_API_KEY}
       - MCP_BRAVE_URL=http://mcp-brave:8080
       - MCP_FETCH_URL=http://mcp-fetch:8000

backend/docker-compose.yml

@@ -6,7 +6,7 @@ services:
     build: ./router
     ports:
       - "0.0.0.0:8000:8000"  # Bind to all interfaces
-      - "0.0.0.0:8443:8443"# HTTPS port (uncomment if using SSL)
+      - "0.0.0.0:8443:8443" # HTTPS port (uncomment if using SSL)
     environment:
       - LOG_LEVEL=DEBUG
       - HARMONY_REASONING_EFFORT=low
@@ -17,12 +17,13 @@ services:
       - PYTHONUNBUFFERED=1
       - PYTHONDONTWRITEBYTECODE=1
       - WATCHDOG_POLLING=true
-      - REMOTE_INFERENCE_URL=https://api.studio.nebius.com
-      - RATING_INFERENCE_URL=https://api.openai.com
-      - USE_REMOTE_INFERENCE=${USE_REMOTE_INFERENCE}
-      - OPENAI_KEY=${OPENAI_KEY}
+      - RATING_INFERENCE_URL=${RATING_INFERENCE_URL}
+      - RATING_INFERENCE_KEY=${RATING_INFERENCE_KEY}
+      - RATING_INFERENCE_MODEL=${RATING_INFERENCE_MODEL}
       - REMOTE_INFERENCE_KEY=${REMOTE_INFERENCE_KEY}
-      - USE_REMOTE_INFERENCE=true
+      - USE_REMOTE_INFERENCE=${USE_REMOTE_INFERENCE}
+      - REMOTE_INFERENCE_MODEL=${REMOTE_INFERENCE_MODEL}
+      - REMOTE_INFERENCE_URL=${REMOTE_INFERENCE_URL}
       - BRAVE_API_KEY=${BRAVE_API_KEY}
       - MCP_BRAVE_URL=http://mcp-brave:8080
 
@@ -63,7 +64,7 @@ services:
     build: ./router
     ports:
       - "8000:8000"  # Bind to all interfaces
-      - "8443:8443"# HTTPS port (uncomment if using SSL)
+      - "8443:8443" # HTTPS port (uncomment if using SSL)
     environment:
       - LOG_LEVEL=DEBUG
       - HARMONY_REASONING_EFFORT=low
@@ -75,11 +76,6 @@ services:
       - PYTHONDONTWRITEBYTECODE=1
       - WATCHDOG_POLLING=true
       - MCP_BRAVE_URL=http://mcp-brave:8080
-      - OPENAI_URL=https://api.openai.com
-      - USE_REMOTE_INFERENCE=${USE_REMOTE_INFERENCE}
-      - OPENAI_KEY=${OPENAI_KEY}
-      - USE_REMOTE_INFERENCE=true
-      - BRAVE_API_KEY=${BRAVE_API_KEY}
 
     volumes:
       # Mount source code for live reloading
@@ -132,9 +128,13 @@ services:
       - DISABLE_PREMIUM_CHECK=true
       - WATCHDOG_POLLING=true
       - MCP_BRAVE_URL=http://mcp-brave:8080
-      - OPENAI_URL=https://api.openai.com
-      - USE_REMOTE_INFERENCE=false
-      - OPENAI_KEY=${OPENAI_KEY}
+      - RATING_INFERENCE_URL=${RATING_INFERENCE_URL}
+      - RATING_INFERENCE_KEY=${RATING_INFERENCE_KEY}
+      - RATING_INFERENCE_MODEL=${RATING_INFERENCE_MODEL}
+      - REMOTE_INFERENCE_KEY=${REMOTE_INFERENCE_KEY}
+      - USE_REMOTE_INFERENCE=${USE_REMOTE_INFERENCE}
+      - REMOTE_INFERENCE_MODEL=${REMOTE_INFERENCE_MODEL}
+      - REMOTE_INFERENCE_URL=${REMOTE_INFERENCE_URL}
       - BRAVE_API_KEY=${BRAVE_API_KEY}
       - MCP_FETCH_URL=http://mcp-fetch:8000
 

backend/env.example

@@ -1,7 +1,12 @@
-# OpenAI Configuration
-OPENAI_KEY=your-openai-api-key-here
-
-# Brave api key
+# Inference configuration
+RATING_INFERENCE_URL=${RATING_INFERENCE_URL}
+RATING_INFERENCE_KEY=${RATING_INFERENCE_KEY}
+RATING_INFERENCE_MODEL=${RATING_INFERENCE_MODEL}
+REMOTE_INFERENCE_KEY=${REMOTE_INFERENCE_KEY}
+USE_REMOTE_INFERENCE=${USE_REMOTE_INFERENCE}
+REMOTE_INFERENCE_MODEL=${REMOTE_INFERENCE_MODEL}
+REMOTE_INFERENCE_URL=${REMOTE_INFERENCE_U
+# Brapi key
 BRAVE_API_KEY=your-brave-api-key-here
 
 # Service URLs
@@ -16,7 +21,6 @@ REASONING_EFFORT=low
 API_HOST=0.0.0.0
 API_PORT=8000
 
-
 # Timeouts
 INFERENCE_TIMEOUT=60
 EMBEDDINGS_TIMEOUT=60
@@ -27,8 +31,7 @@ LOG_LEVEL=INFO
 # Tool Calling Configuration
 ENABLE_TOOL_CALLS=true
 
+# Remote Inference
 USE_REMOTE_INFERENCE=true
 
-ENABLE_TOOL_CALLS=false
-