Image Generation tool for Presentation Generator (PoC)

.gitignore

@@ -176,3 +176,6 @@ compose*.yaml
 # vscode
 .vscode/
 
+# Google Cloud Service Account credentials
+marvel-ai-backend-credentials.json
+

Dockerfile

@@ -10,6 +10,9 @@ COPY requirements.txt /code/requirements.txt
 
 RUN pip install --no-cache-dir -r /code/requirements.txt
 
+# Copy credentials file
+COPY marvel-ai-backend-credentials.json /code/marvel-ai-backend-credentials.json
+
 COPY ./app /code/app
 
 # Local development key set

app/services/schemas.py

@@ -175,4 +175,7 @@ class SlideGeneratorInput(BaseModel):
     slides_titles: List[str]
     instructional_level: str
     topic: str
-    lang: Optional[str] = "en"
+    lang: Optional[str] = "en"
+
+class SlideImageGeneratorInput(BaseModel):
+    slides: List[dict]

app/tools/presentation_generator_updated/slide_generator/tools.py

@@ -102,12 +102,12 @@ def compile_context(self):
         )
         chain = prompt | self.model | self.parser
 
-        logger.info(f"Chain compilation complete")
+        logger.info("Chain compilation complete")
 
         return chain
 
     def generate_slides(self):
-        logger.info(f"Creating the Outlines for the Presentation") 
+        logger.info("Creating the Outlines for the Presentation") 
         chain = self.compile_context() 
 
         input_parameters = {
@@ -126,7 +126,7 @@ def generate_slides(self):
         logger.info(f"Response validation: {validation_results}")
 
         if not validation_results["valid"]:
-            logger.warning(f"Generated content may not fully match the requested topic")
+            logger.warning("Generated content may not fully match the requested topic")
         return response
 
 class Slide(BaseModel):

app/tools/presentation_generator_updated/slide_image_generator/core.py

@@ -0,0 +1,34 @@
+from typing import List, Dict
+from app.services.logger import setup_logger
+from app.tools.presentation_generator_updated.slide_image_generator.tools import SlideImageGenerator
+
+logger = setup_logger(__name__)
+
+def executor(slides: List[dict], verbose: bool = False) -> dict:
+    """
+    Execute the slide image generation process.
+
+    Args:
+        slides (List[dict]): List of slide dictionaries from the slide generator
+        verbose (bool): Enable verbose logging
+
+    Returns:
+        dict: Dictionary containing processed slides with generated images
+    """
+    try:
+        if verbose: 
+            logger.info("Starting slide image generation process")
+        generator = SlideImageGenerator(slides=slides, verbose=verbose)
+
+        if verbose: 
+            logger.info("Generating slides images")
+        result = generator.generate_slides()
+
+        if verbose: 
+            logger.info("Slides images generated successfully")
+        return result
+
+    except Exception as e:
+        error_message = f"Error in slide image generation: {str(e)}"
+        logger.error(error_message)
+        raise ValueError(error_message)

app/tools/presentation_generator_updated/slide_image_generator/metadata.json

@@ -0,0 +1,21 @@
+{
+    "inputs": [
+        {
+            "label": "Slides",
+            "name": "slides",
+            "type": "list",
+            "item_type": "object",
+            "properties": {
+                "title": {
+                    "type": "string"
+                },
+                "template": {
+                    "type": "string"
+                },
+                "content": {
+                    "type": "any"
+                }
+            }
+        }
+    ]
+}

app/tools/presentation_generator_updated/slide_image_generator/tests/test_service_account.py

@@ -0,0 +1,121 @@
+# test_service_account.py
+import os
+import io
+from pathlib import Path
+from PIL import Image
+from google.cloud import storage
+import vertexai
+from vertexai.preview.vision_models import ImageGenerationModel
+from dotenv import load_dotenv
+
+# Find the project root directory and load environment variables
+current_file = Path(__file__)
+project_root = current_file.parent
+# Navigate up until we find the project root (where the app directory is)
+while not (project_root / 'app').exists() and project_root != project_root.parent:
+    project_root = project_root.parent
+
+# Look for .env in common locations
+env_paths = [
+    project_root / 'app' / '.env',  # app/.env
+    project_root / '.env',          # .env in project root
+]
+
+# Try to load from the first .env file found
+env_loaded = False
+for env_path in env_paths:
+    if env_path.exists():
+        print(f"Loading environment from: {env_path}")
+        load_dotenv(env_path)
+        env_loaded = True
+        break
+
+if not env_loaded:
+    print("Warning: No .env file found. Using environment variables as is.")
+
+def test_service_account():
+    print("\n=== Testing Google Cloud Service Account ===\n")
+
+    # 1. Find and verify credentials
+    print("Step 1: Finding credentials file")
+    creds_path = os.getenv('GOOGLE_APPLICATION_CREDENTIALS')
+
+    # Make relative path absolute if needed
+    if creds_path and not os.path.isabs(creds_path):
+        creds_path = os.path.join(project_root, creds_path)
+
+    # Use default location if not specified
+    if not creds_path:
+        creds_path = project_root / 'marvel-ai-backend-credentials.json' # as per your credentials file name
+
+    # Check if file exists
+    if os.path.exists(creds_path):
+        print(f"✅ Credentials file found at: {creds_path}")
+        os.environ['GOOGLE_APPLICATION_CREDENTIALS'] = str(creds_path)
+    else:
+        print(f"❌ Credentials file not found at: {creds_path}")
+        return
+
+    # 2. Generate a test image with Vertex AI
+    print("\nStep 2: Generating test image with Vertex AI")
+    project_id = os.getenv('PROJECT_ID')
+    if not project_id:
+        print("❌ PROJECT_ID not set in environment")
+        return
+
+    print(f"Initializing Vertex AI with project: {project_id}")
+    vertexai.init(project=project_id)
+
+    # Generate a simple test image
+    print("Generating image...")
+    image_model = ImageGenerationModel.from_pretrained("imagen-3.0-fast-generate-001")
+    prompt = "A simple blue circle on a white background, abstract, no text"
+
+    try:
+        response = image_model.generate_images(
+            prompt=prompt,
+            number_of_images=1,
+        )
+
+        if response and hasattr(response[0], '_image_bytes'):
+            image_bytes = response[0]._image_bytes
+            print(f"✅ Successfully generated image ({len(image_bytes)} bytes)")
+
+            # Save image locally for verification
+            with open("test_image.png", "wb") as f:
+                f.write(image_bytes)
+            print("✅ Saved test image to test_image.png")
+        else:
+            print("❌ Failed to generate image")
+            return
+    except Exception as e:
+        print(f"❌ Image generation error: {str(e)}")
+        return
+
+    # 3. Store the image in Cloud Storage
+    print("\nStep 3: Storing image in Cloud Storage")
+    bucket_name = os.getenv('SLIDE_IMAGES_BUCKET_NAME', 'presentation-slide-images')
+    print(f"Using bucket: {bucket_name}")
+
+    try:
+        storage_client = storage.Client()
+        bucket = storage_client.bucket(bucket_name)
+
+        # Upload the generated image
+        blob = bucket.blob("test_image.png")
+        blob.upload_from_string(image_bytes, content_type="image/png")
+        print(f"✅ Successfully uploaded image to: gs://{bucket_name}/test_image.png")
+
+        # Make it publicly accessible and get URL
+        blob.make_public()
+        print(f"✅ Public URL: {blob.public_url}")
+
+        # Clean up (optional)
+        # blob.delete()
+        # print("✅ Deleted test image from bucket")
+
+    except Exception as e:
+        print(f"❌ Storage error: {str(e)}")
+
+if __name__ == "__main__":
+    test_service_account()