Update app.py

app.py

@@ -5,14 +5,19 @@ import base64
 from PIL import Image
 import io
 import json
+import copy
+import secrets
+from pathlib import Path
+import re
 
+# Funções auxiliares
 def resize_image(image_path, max_size=(800, 800), quality=85):
     with Image.open(image_path) as img:
         img.thumbnail(max_size, Image.Resampling.LANCZOS)
         buffer = io.BytesIO()
         img.save(buffer, format="JPEG", quality=quality)
         return buffer.getvalue()
-        
+
 def filepath_to_base64(image_path):
     img_bytes = resize_image(image_path)
     img_base64 = base64.b64encode(img_bytes)
@@ -20,13 +25,8 @@ def filepath_to_base64(image_path):
 
 api_key = os.getenv('API_KEY')
 
+# Função principal da API
 def call_fuyu_8b_api(image_path, content, temperature=0.2, top_p=0.7, max_tokens=1024):
-    print(f"Caminho da imagem recebida: {image_path}")
-    print(f"Conteúdo: {content}")
-    print(f"Temperatura: {temperature}")
-    print(f"Top P: {top_p}")
-    print(f"Max Tokens: {max_tokens}")
-
     image_base64 = filepath_to_base64(image_path)
     invoke_url = "https://api.nvcf.nvidia.com/v2/nvcf/pexec/functions/9f757064-657f-4c85-abd7-37a7a9b6ee11"
     headers = {
@@ -56,39 +56,52 @@ def call_fuyu_8b_api(image_path, content, temperature=0.2, top_p=0.7, max_tokens
             print(error_details)
         except ValueError:
             print(response.text)
+        return "Error in API call"
     else:
         response_text = ""
         for line in response.iter_lines():
             if line:
-                try:
-                    # Decode the line from bytes to string
-                    decoded_line = line.decode('utf-8')
-        
-                    # Remove the "data: " prefix
-                    if decoded_line.startswith('data: '):
-                        json_str = decoded_line[6:]  # Remove the first 6 characters ('data: ')
-                        json_line = json.loads(json_str)
-        
-                        # Assuming the structure is consistent with the examples you provided.
-                        content_parts = json_line.get("choices", [{}])[0].get("delta", {}).get("content", "")
-                        response_text += content_parts
-                    else:
-                        print(f"Unexpected line format: {decoded_line}")
-                except json.JSONDecodeError as e:
-                    print(f"Error decoding JSON from response line: {e}")
-                    print(f"Faulty line: {line}")
-
+                decoded_line = line.decode('utf-8')
+                if decoded_line.startswith('data: '):
+                    json_str = decoded_line[6:]
+                    json_line = json.loads(json_str)
+                    content_parts = json_line.get("choices", [{}])[0].get("delta", {}).get("content", "")
+                    response_text += content_parts
         return response_text
-content_input = gr.Textbox(lines=2, placeholder="Enter your content here...", label="Content")
-image_input = gr.Image(type="filepath", label="Upload Image")
-temperature_input = gr.Slider(minimum=0, maximum=1, step=0.01, value=0.2, label="Temperature")
-top_p_input = gr.Slider(minimum=0, maximum=1, step=0.01, value=0.7, label="Top P")
-max_tokens_input = gr.Slider(minimum=1, maximum=1024, step=1, value=1024, label="Max Tokens")
 
-iface = gr.Interface(fn=call_fuyu_8b_api,
-                     inputs=[image_input, content_input, temperature_input, top_p_input, max_tokens_input],
-                     outputs="text",
-                     title="Fuyu-8B API Explorer",
-                     description="Explore the capabilities of Fuyu-8B multi-modal transformer.")
+# Interface Gradio
+def update_chatbot_with_response(chatbot, response):
+    chatbot.append(("Assistant", response))
+
+def submit_response(chatbot, task_history, image_input, content_input, temperature_input, top_p_input, max_tokens_input):
+    image_path = image_input if image_input else "No image provided"
+    response = call_fuyu_8b_api(image_path, content_input, temperature_input, top_p_input, max_tokens_input)
+    update_chatbot_with_response(chatbot, response)
+    return chatbot, task_history
+
+def reset_state(task_history):
+    task_history.clear()
+    return [], task_history
+
+css = '''
+.gradio-container{max-width:800px !important}
+'''
+
+with gr.Blocks(css=css) as demo:
+    gr.Markdown("# Fuyu-8B API Explorer with Enhanced Features")
+    chatbot = gr.Chatbot(label='Chatbot', elem_classes="control-height", height=520)
+    task_history = gr.State([])
+    with gr.Row():
+        addfile_btn = gr.File(label="📁 Upload Image", file_types=["jpg", "png"])
+        submit_btn = gr.Button("🚀 Submit")
+        regen_btn = gr.Button("🤔 Regenerate")
+        empty_bin = gr.Button("🧹 Clear History")
+    with gr.Accordion("Advanced Settings"):
+        content_input = gr.Textbox(lines=2, placeholder="Enter your content here...", label="Content")
+        temperature_input = gr.Slider(minimum=0, maximum=1, step=0.01, value=0.2, label="Temperature")
+        top_p_input = gr.Slider(minimum=0, maximum=1, step=0.01, value=0.7, label="Top P")
+        max_tokens_input = gr.Slider(minimum=1, maximum=1024, step=1, value=1024, label="Max Tokens")
+    submit_btn.click(submit_response, [chatbot, task_history, addfile_btn, content_input, temperature_input, top_p_input, max_tokens_input], [chatbot, task_history])
+    empty_bin.click(reset_state, [task_history], [chatbot, task_history])
 
-iface.launch()
+demo.launch()