Build Real World AI Applications with Gemini and Imagen

GSP — Build a Multi-Modal GenAI Application: Challenge Lab

Estimated time: 1 hour

# 🚀 Gemini Enterprise Agent Ready (GEAR) Challenge Lab: Image Generation and Analysis > ⚠️ **Disclaimer:** This is an independent, community-made walkthrough created for educational purposes, hands-on practice, and Google Cloud certification preparation. This guide is designed to help learners understand generative AI, Vertex AI, image generation, multimodal analysis, and practical Python workflows using the unified `google-genai` SDK. Always attempt the lab yourself first and follow Google Cl

# ==============================================================================
# ORBIT OF OPS - GEAR CHALLENGE LAB (COMMAND 1 OF 1)
# ==============================================================================
GREEN='\e[1;32m'
CYAN='\e[1;36m'
YELLOW='\e[1;33m'
BLUE='\e[1;34m'
MAGENTA='\e[1;35m'
RED='\e[1;31m'
RESET='\e[0m'
BOLD='\e[1m'

clear
echo -e "${CYAN}${BOLD}"
cat << "EOF"
  ____       _     _ _            __    ___            
 / __ \     | |   (_) |          / _|  / _ \           
| |  | |_ __| |__  _| |_   ___  | |_  | | | |_ __  ___ 
| |  | | '__| '_ \| | __| / _ \ |  _| | | | | '_ \/ __|
| |__| | |  | |_) | | |_ | (_) || |   | |_| | |_) \__ \
 \____/|_|  |_.__/|_|\__| \___/ |_|    \___/| .__/|___/
                                            | |        
                                            |_|        
EOF
echo -e "${RESET}"
echo -e "${MAGENTA}${BOLD}>>> ORBIT OF OPS: GEMINI GEAR CHALLENGE INITIATED <<<${RESET}\n"

# ==============================================================================
# PRE-FLIGHT CHECKS & VARIABLES
# ==============================================================================
echo -e "${BOLD}${YELLOW}[Orbit of Ops] Auto-fetching Project...${RESET}"
export PROJECT_ID=$(gcloud config get-value project 2>/dev/null)
if [[ -z "$PROJECT_ID" ]]; then export PROJECT_ID=$DEVSHELL_PROJECT_ID; fi
echo -e "✅ Project ID: ${GREEN}$PROJECT_ID${RESET}\n"

echo -e "${YELLOW}${BOLD}--- REQUIRED LAB INPUTS ---${RESET}"
read -p "$(echo -e "${CYAN}${BOLD}1. Enter the Lab Region (e.g., us-central1): ${RESET}") " REGION
read -p "$(echo -e "${MAGENTA}${BOLD}2. Enter the flash-image-model-id (e.g., imagen-3.0-generate-001): ${RESET}") " IMAGE_MODEL_ID
read -p "$(echo -e "${BLUE}${BOLD}3. Enter the model-id (e.g., gemini-1.5-flash): ${RESET}") " TEXT_MODEL_ID

export REGION
export IMAGE_MODEL_ID
export TEXT_MODEL_ID

# ==============================================================================
# STEP 1: PROVISION ENVIRONMENT
# ==============================================================================
echo -e "\n${BLUE}${BOLD}[Orbit of Ops] Installing Required Python Packages...${RESET}"
pip install google-genai pillow --quiet --disable-pip-version-check

# ==============================================================================
# STEP 2: WRITE PYTHON SCRIPT
# ==============================================================================
echo -e "${CYAN}${BOLD}[Orbit of Ops] Generating Python Pipeline...${RESET}"
cat << 'EOF' > solution.py
import os
from google import genai
from google.genai import types
from PIL import Image

# Retrieve environment variables
PROJECT_ID = os.environ.get("PROJECT_ID")
REGION = os.environ.get("REGION")
IMAGE_MODEL_ID = os.environ.get("IMAGE_MODEL_ID")
TEXT_MODEL_ID = os.environ.get("TEXT_MODEL_ID")

print(f"\nInitializing Vertex AI client in {REGION} for {PROJECT_ID}...")
client = genai.Client(vertexai=True, project=PROJECT_ID, location=REGION)

# ==========================================
# Task 1: Generate Bouquet Image
# ==========================================
print("\n[Task 1] Generating bouquet image using model:", IMAGE_MODEL_ID)
image_prompt = "Create an image containing a bouquet of 2 sunflowers and 3 roses."
image_path = "bouquet.jpg"

try:
    image_result = client.models.generate_images(
        model=IMAGE_MODEL_ID,
        prompt=image_prompt,
        config=types.GenerateImagesConfig(
            number_of_images=1,
            output_mime_type="image/jpeg"
        )
    )

    for generated_image in image_result.generated_images:
        generated_image.image.save(image_path)
        print(f"✅ Image successfully generated and saved locally to {image_path}")

except Exception as e:
    print(f"❌ Error during image generation: {e}")
    exit(1)


# ==========================================
# Task 2: Analyze Bouquet Image (Streamed)
# ==========================================
print(f"\n[Task 2] Analyzing image using model: {TEXT_MODEL_ID}")

def analyze_bouquet_image(image_path):
    img = Image.open(image_path)
    text_prompt = "Generate birthday wishes inspired by the bouquet image."
    
    print("Initiating streaming request...")
    response = client.models.generate_content_stream(
        model=TEXT_MODEL_ID,
        contents=[img, text_prompt]
    )
    
    print("Writing stream to birthday_wishes.txt...\n")
    print("-" * 40)
    with open("birthday_wishes.txt", "w") as f:
        for chunk in response:
            f.write(chunk.text)
            print(chunk.text, end="", flush=True)
    print("\n" + "-" * 40)
    print("✅ Successfully streamed and saved to birthday_wishes.txt")

try:
    analyze_bouquet_image(image_path)
except Exception as e:
    print(f"❌ Error during image analysis: {e}")
    exit(1)
EOF

# ==============================================================================
# STEP 3: EXECUTE SCRIPT
# ==============================================================================
echo -e "${YELLOW}${BOLD}[Orbit of Ops] Executing AI Pipeline...${RESET}"
python3 solution.py

echo -e "\n${GREEN}${BOLD}🎉 COMMAND 1 COMPLETE!${RESET}"
echo -e "${CYAN}${BOLD}You can now safely click 'Check my progress' on Task 1 and Task 2 in your lab manual.${RESET}"