databuild/tests/end_to_end/podcast_simple_test.sh

#!/bin/bash

# Simple end-to-end test for podcast reviews
set -euo pipefail

CLI_BUILD="${1:-}"
SERVICE_BINARY="${2:-}"

if [[ -z "$CLI_BUILD" ]] || [[ -z "$SERVICE_BINARY" ]]; then
    echo "Usage: $0 <cli_build_binary> <service_binary>"
    exit 1
fi

echo "[INFO] Testing CLI build for podcast reviews..."
export DATABUILD_BUILD_EVENT_LOG="sqlite:///tmp/podcast_test_cli.db"
rm -f /tmp/podcast_test_cli.db

# Test CLI build with a simple partition
if ! "$CLI_BUILD" "reviews/date=2020-01-01" > /tmp/podcast_cli_output.log 2>&1; then
    echo "[ERROR] CLI build failed"
    cat /tmp/podcast_cli_output.log
    exit 1
fi

echo "[INFO] CLI build succeeded"

# Count events in CLI database
if [[ -f /tmp/podcast_test_cli.db ]]; then
    CLI_EVENTS=$(sqlite3 /tmp/podcast_test_cli.db "SELECT COUNT(*) FROM build_events;" 2>/dev/null || echo "0")
    echo "[INFO] CLI generated $CLI_EVENTS events"
else
    echo "[ERROR] CLI database not created"
    exit 1
fi

echo "[INFO] Testing Service build for podcast reviews..."
# The service uses a hardcoded database path
SERVICE_DB_PATH="/tmp/podcast_reviews_graph_service.db"
rm -f "$SERVICE_DB_PATH"

# Start service
SERVICE_PORT=58082
"$SERVICE_BINARY" --port="$SERVICE_PORT" --host="127.0.0.1" > /tmp/podcast_service.log 2>&1 &
SERVICE_PID=$!

# Cleanup service on exit
trap "kill $SERVICE_PID 2>/dev/null || true; wait $SERVICE_PID 2>/dev/null || true" EXIT

# Wait for service to start and test health
sleep 3
for i in {1..10}; do
    if curl -s "http://127.0.0.1:$SERVICE_PORT/api/v1/builds" > /dev/null 2>&1; then
        echo "[INFO] Service is healthy"
        break
    fi
    if [[ $i -eq 10 ]]; then
        echo "[ERROR] Service health check failed"
        cat /tmp/podcast_service.log
        exit 1
    fi
    sleep 1
done

# Make build request
BUILD_RESPONSE=$(curl -s -X POST \
    -H "Content-Type: application/json" \
    -d '{"partitions": ["reviews/date=2020-01-01"]}' \
    "http://127.0.0.1:$SERVICE_PORT/api/v1/builds")

BUILD_ID=$(echo "$BUILD_RESPONSE" | jq -r '.build_request_id' 2>/dev/null || echo "")

if [[ -z "$BUILD_ID" || "$BUILD_ID" == "null" ]]; then
    echo "[ERROR] Failed to get build ID: $BUILD_RESPONSE"
    exit 1
fi

echo "[INFO] Created build request: $BUILD_ID"

# Wait for build completion
for i in {1..60}; do
    STATUS_RESPONSE=$(curl -s "http://127.0.0.1:$SERVICE_PORT/api/v1/builds/$BUILD_ID")
    STATUS=$(echo "$STATUS_RESPONSE" | jq -r '.status' 2>/dev/null || echo "UNKNOWN")
    
    case "$STATUS" in
        "completed"|"COMPLETED")
            echo "[INFO] Service build completed"
            break
            ;;
        "failed"|"FAILED")
            echo "[ERROR] Service build failed: $STATUS_RESPONSE"
            exit 1
            ;;
        "running"|"RUNNING"|"pending"|"PENDING"|"planning"|"PLANNING")
            echo "[INFO] Build status: $STATUS"
            sleep 2
            ;;
        *)
            echo "[WARN] Unknown status: $STATUS"
            sleep 2
            ;;
    esac
    
    if [[ $i -eq 60 ]]; then
        echo "[ERROR] Build did not complete within 120 seconds"
        exit 1
    fi
done

# Count events in Service database
if [[ -f "$SERVICE_DB_PATH" ]]; then
    SERVICE_EVENTS=$(sqlite3 "$SERVICE_DB_PATH" "SELECT COUNT(*) FROM build_events;" 2>/dev/null || echo "0")
    echo "[INFO] Service generated $SERVICE_EVENTS events"
else
    echo "[ERROR] Service database not created"
    exit 1
fi

# Compare event counts with detailed analysis
if [[ "$CLI_EVENTS" -gt 0 ]] && [[ "$SERVICE_EVENTS" -gt 0 ]]; then
    echo "[INFO] Both CLI and Service generated events successfully"
    echo "[INFO] CLI: $CLI_EVENTS events, Service: $SERVICE_EVENTS events"
    
    # Detailed event comparison
    echo "[INFO] Analyzing event distribution..."
    
    CLI_JOB_EVENTS=$(sqlite3 /tmp/podcast_test_cli.db "SELECT COUNT(*) FROM build_events WHERE event_type = 'job';" 2>/dev/null || echo "0")
    CLI_PARTITION_EVENTS=$(sqlite3 /tmp/podcast_test_cli.db "SELECT COUNT(*) FROM build_events WHERE event_type = 'partition';" 2>/dev/null || echo "0")
    CLI_REQUEST_EVENTS=$(sqlite3 /tmp/podcast_test_cli.db "SELECT COUNT(*) FROM build_events WHERE event_type = 'build_request';" 2>/dev/null || echo "0")
    
    SERVICE_JOB_EVENTS=$(sqlite3 "$SERVICE_DB_PATH" "SELECT COUNT(*) FROM build_events WHERE event_type = 'job';" 2>/dev/null || echo "0")
    SERVICE_PARTITION_EVENTS=$(sqlite3 "$SERVICE_DB_PATH" "SELECT COUNT(*) FROM build_events WHERE event_type = 'partition';" 2>/dev/null || echo "0")
    SERVICE_REQUEST_EVENTS=$(sqlite3 "$SERVICE_DB_PATH" "SELECT COUNT(*) FROM build_events WHERE event_type = 'build_request';" 2>/dev/null || echo "0")
    
    echo "[INFO] Event breakdown:"
    echo "[INFO]   Job events:       CLI=$CLI_JOB_EVENTS, Service=$SERVICE_JOB_EVENTS"
    echo "[INFO]   Partition events: CLI=$CLI_PARTITION_EVENTS, Service=$SERVICE_PARTITION_EVENTS"  
    echo "[INFO]   Request events:   CLI=$CLI_REQUEST_EVENTS, Service=$SERVICE_REQUEST_EVENTS"
    
    # Validate core events are identical (job, partition, and request events should all match now)
    if [[ "$CLI_JOB_EVENTS" -eq "$SERVICE_JOB_EVENTS" ]] && [[ "$CLI_PARTITION_EVENTS" -eq "$SERVICE_PARTITION_EVENTS" ]] && [[ "$CLI_REQUEST_EVENTS" -eq "$SERVICE_REQUEST_EVENTS" ]]; then
        echo "[INFO] ✅ All build events (job, partition, and request) are identical"
    else
        echo "[ERROR] ❌ Build events differ between CLI and Service - this indicates a problem"
        exit 1
    fi
    
    # Validate total event counts are identical  
    if [[ "$CLI_EVENTS" -eq "$SERVICE_EVENTS" ]]; then
        echo "[INFO] ✅ Total event counts are identical: $CLI_EVENTS events each"
    else
        echo "[ERROR] ❌ Total event counts differ: CLI=$CLI_EVENTS, Service=$SERVICE_EVENTS"
        exit 1
    fi
    
else
    echo "[ERROR] One or both builds generated no events"
    exit 1
fi

echo "[INFO] Podcast reviews simple end-to-end test completed successfully!"