Update app.py

app.py

@@ -1,14 +1,20 @@
+import threading
+import time
 from smolagents import CodeAgent, DuckDuckGoSearchTool, HfApiModel, tool
 import datetime
 import pytz
 import yaml
 import requests
-import time
 import base64
 from io import BytesIO
 from PIL import Image
 from tools.final_answer import FinalAnswerTool
 from Gradio_UI import GradioUI
+from huggingface_hub import InferenceClient
+
+# --- Set up the Inference Client for Image Generation ---
+# This client calls the Studio Ghibli Image Generator V2 model.
+image_client = InferenceClient(model_id="artificialguybr/StudioGhibli.Redmond-V2")
 
 # --- Define Tools ---
 
@@ -61,7 +67,7 @@ def probe_story_preferences(likes: str, dislikes: str, favorite_theme: str) -> s
         favorite_theme: Your favored narrative theme (e.g., adventure, mystery).
     
     Returns:
-        A summary that shapes your personalized tale.
+        A summary that helps weave your personalized tale.
     """
     return f"Splendid! You delight in {likes} 🌸, wish to steer clear of {dislikes} 🚫, and are enchanted by {favorite_theme} themes. Let us craft your story accordingly."
 
@@ -71,62 +77,36 @@ def generate_studio_ghibli_image(prompt: str) -> str:
     Generates a Studio Ghibli style image based on the provided prompt.
     
     Args:
-        prompt: The text prompt describing the desired image.
+        prompt: The descriptive prompt for the desired image.
     
     Returns:
-        A base64 encoded PNG image string representing the generated image.
+        A base64-encoded PNG string representing the generated image.
     """
-    repo = "artificialguybr/StudioGhibli.Redmond-V2"
-    trigger_word = "Studio Ghibli, StdGBRedmAF"
-    api_url = f"https://api-inference.huggingface.co/models/{repo}"
-    # Optionally, set up your Hugging Face token in the environment and uncomment below:
-    # token = os.getenv("API_TOKEN")
-    headers = {
-        # "Authorization": f"Bearer {token}"
-    }
-    full_prompt = f"{prompt} {trigger_word}"
-    payload = {
-        "inputs": full_prompt,
-        "parameters": {
-            "negative_prompt": "(worst quality, low quality, normal quality, lowres, low details, oversaturated, undersaturated, overexposed, underexposed, grayscale, bw, bad photo, bad photography, bad art:1.4), (watermark, signature, text font, username, error, logo, words, letters, digits, autograph, trademark, name:1.2), (blur, blurry, grainy), morbid, ugly, asymmetrical, mutated malformed, mutilated, poorly lit, bad shadow, draft, cropped, out of frame, cut off, censored, jpeg artifacts, out of focus, glitch, duplicate, (airbrushed, cartoon, anime, semi-realistic, cgi, render, blender, digital art, manga, amateur:1.3), (3D ,3D Game, 3D Game Scene, 3D Character:1.1), (bad hands, bad anatomy, bad body, bad face, bad teeth, bad arms, bad legs, deformities:1.3)",
-            "num_inference_steps": 30,
-            "scheduler": "DPMSolverMultistepScheduler"
-        },
-    }
-    error_count = 0
-    while True:
-        response = requests.post(api_url, headers=headers, json=payload)
-        if response.status_code == 200:
-            img = Image.open(BytesIO(response.content))
-            buffered = BytesIO()
-            img.save(buffered, format="PNG")
-            img_str = base64.b64encode(buffered.getvalue()).decode("utf-8")
-            return img_str
-        elif response.status_code == 503:
-            time.sleep(1)
-        elif response.status_code == 500 and error_count < 5:
-            time.sleep(1)
-            error_count += 1
-        else:
-            raise Exception(f"API Error: {response.status_code}")
+    try:
+        # Call the image-generation API with our prompt and desired parameters.
+        result = image_client.image_generation(prompt, parameters={"width": 1024, "height": 1024, "prompt": prompt})
+        # Assume result is a base64 string or a URL; here we return it directly.
+        return result
+    except Exception as e:
+        return f"Error generating image: {str(e)}"
 
 @tool
 def offer_interactive_options(scene: str) -> str:
     """
-    Provides interactive options for the next step in your adventure based on the scene.
+    Provides interactive options based on the current scene description.
     
     Args:
-        scene: A detailed description of the current scene.
+        scene: The detailed description of the current scene.
     
     Returns:
-        A string listing numbered interactive options.
+        A string listing interactive options.
     """
     return (
         "Options:\n"
         "1. Follow the mysterious figure in the distance.\n"
         "2. Wander deeper into the enchanted landscape.\n"
         "3. Approach a peculiar, glowing object.\n"
-        "4. Ask for more secrets of the scene.\n"
+        "4. Inquire for more secrets of the scene.\n"
         "Please enter the number of your choice."
     )
 
@@ -176,7 +156,7 @@ agent = CodeAgent(
     name="Ghibili Interactive Tale",
     description=(
         "Ghibili Interactive Tale is your creative partner on a magical journey. "
-        "We will co-author a story as enchanting as a classic tale, with every scene painted in Studio Ghibli style. "
+        "Together, we'll co-author a story as enchanting as a classic tale, with every scene painted in Studio Ghibli style. "
         "I begin by reflecting on your current time and mood, ask for your location to capture the weather, "
         "and then weave a narrative filled with vivid imagery and interactive choices. Let's create magic together!"
     ),
@@ -193,7 +173,28 @@ initial_opener = (
     "What kind of tale are you yearning for today—a whimsical fairy tale, a mysterious quest, or a serene escape? "
     "Your story awaits, and I am excited to walk beside you every step of the way. 📖✨"
 )
-agent.conversation = [{"role": "Storyteller", "content": initial_opener}]
+agent.conversation = [{"role": "assistant", "content": initial_opener}]
+
+# --- Proactive Story Loop ---
+# This loop proactively generates narrative segments if the user is idle.
+def proactive_story_loop(agent, interval=15, steps=5):
+    for i in range(steps):
+        # Simulate generation of the next narrative segment.
+        # Here, we assume that `agent.generate_next_message()` generates a new story segment based on the current conversation.
+        try:
+            new_segment = agent.generate_next_message()
+        except Exception as e:
+            new_segment = f"(Error generating next message: {e})"
+        if new_segment:
+            agent.conversation.append({"role": "assistant", "content": new_segment})
+        time.sleep(interval)
+    # After the narrative loop, generate a final image representing the culmination of the story.
+    image_prompt = "A breathtaking final scene that encapsulates the magical journey we have created together, in vivid Studio Ghibli style."
+    final_image = generate_studio_ghibli_image(image_prompt)
+    agent.conversation.append({"role": "assistant", "content": f"Here is the final image of our adventure: {final_image}"})
+
+# Start the proactive story loop in a background thread.
+threading.Thread(target=proactive_story_loop, args=(agent, 15, 5), daemon=True).start()
 
 # --- Launch the Interactive UI ---
 GradioUI(agent).launch()