Update app.py

app.py

@@ -5,6 +5,7 @@ from bs4 import BeautifulSoup
 import torch
 import gradio as gr
 from transformers import AutoModelForSeq2SeqLM, AutoTokenizer
+from huggingface_hub import InferenceClient
 import logging
 
 # Set up logging
@@ -26,6 +27,7 @@ except Exception as e:
     model = None
     tokenizer = None
 
+# Assert to ensure tokenizer is loaded
 assert tokenizer is not None, "Tokenizer failed to load and is None"
 
 # Function to perform a Google search and return the results
@@ -83,15 +85,11 @@ def search(term, num_results=2, lang="en", timeout=5, safe="active", ssl_verify=
 
 # Function to extract visible text from HTML content
 def extract_text_from_webpage(html_content):
-    try:
-        soup = BeautifulSoup(html_content, "html.parser")
-        for tag in soup(["script", "style", "header", "footer", "nav"]):
-            tag.extract()
-        visible_text = soup.get_text(strip=True)
-        return visible_text
-    except Exception as e:
-        logger.error(f"Error extracting text from webpage: {e}")
-        return ""
+    soup = BeautifulSoup(html_content, "html.parser")
+    for tag in soup(["script", "style", "header", "footer", "nav"]):
+        tag.extract()
+    visible_text = soup.get_text(strip=True)
+    return visible_text
 
 # Function to format the prompt for the language model
 def format_prompt(user_prompt, chat_history):
@@ -168,7 +166,7 @@ def model_inference(
         return "Image input not supported in this implementation."
 
 # Define Gradio interface components
-max_new_tokens_slider = gr.Slider(
+max_new_tokens = gr.Slider(
     minimum=2048,
     maximum=16000,
     value=4096,
@@ -176,7 +174,7 @@ max_new_tokens_slider = gr.Slider(
     interactive=True,
     label="Maximum number of new tokens to generate",
 )
-repetition_penalty_slider = gr.Slider(
+repetition_penalty = gr.Slider(
     minimum=0.01,
     maximum=5.0,
     value=1,
@@ -185,7 +183,7 @@ repetition_penalty_slider = gr.Slider(
     label="Repetition penalty",
     info="1.0 is equivalent to no penalty",
 )
-decoding_strategy_radio = gr.Radio(
+decoding_strategy = gr.Radio(
     [
         "Greedy",
         "Top P Sampling",
@@ -195,7 +193,7 @@ decoding_strategy_radio = gr.Radio(
     interactive=True,
     info="Higher values are equivalent to sampling more low-probability tokens.",
 )
-temperature_slider = gr.Slider(
+temperature = gr.Slider(
     minimum=0.0,
     maximum=2.0,
     value=0.5,
@@ -205,7 +203,7 @@ temperature_slider = gr.Slider(
     label="Sampling temperature",
     info="Higher values will produce more diverse outputs.",
 )
-top_p_slider = gr.Slider(
+top_p = gr.Slider(
     minimum=0.01,
     maximum=0.99,
     value=0.9,
@@ -217,14 +215,14 @@ top_p_slider = gr.Slider(
 )
 
 # Create a chatbot interface
-chatbot_interface = gr.Chatbot(
+chatbot = gr.Chatbot(
     label="OpenGPT-4o-Chatty",
     show_copy_button=True,
     likeable=True,
     layout="panel"
 )
 
-# Define Gradio interface function
+# Define Gradio interface
 def chat_interface(user_input, history, web_search, decoding_strategy, temperature, max_new_tokens, repetition_penalty, top_p):
     logger.debug(f"Chat interface called with user_input: {user_input}")
     if isinstance(user_input, str):
@@ -240,27 +238,31 @@ def chat_interface(user_input, history, web_search, decoding_strategy, temperatu
         top_p,
         tokenizer=tokenizer  # Pass tokenizer to model_inference
     )
-    history.append((user_input    ["text"], response))
+    history.append((user_input["text"], response))
+    logger.debug(f"Updated chat history: {history}")
     return history, history
 
-# Create the Gradio interface
+# Create Gradio interface
 interface = gr.Interface(
     fn=chat_interface,
     inputs=[
-        gr.Textbox(placeholder="Type your message here...", lines=1, interactive=True, label="Input"),
-        gr.State(),
-        gr.Checkbox(label="Enable Web Search", value=True, interactive=True),
-        decoding_strategy_radio,
-        temperature_slider,
-        max_new_tokens_slider,
-        repetition_penalty_slider,
-        top_p_slider,
+        gr.Textbox(label="User Input"),
+        gr.State([]),
+        gr.Checkbox(label="Web Search", value=True),
+        decoding_strategy,
+        temperature,
+        max_new_tokens,
+        repetition_penalty,
+        top_p
+    ],
+    outputs=[
+        chatbot,
+        gr.State([])
     ],
-    outputs=[chatbot_interface, gr.State()],
     title="OpenGPT-4o-Chatty",
-    description="A chatbot interface for OpenGPT-4o with web search capabilities.",
-    allow_flagging="never",
+    description="An AI assistant capable of insightful conversations and web search."
 )
 
 if __name__ == "__main__":
-    interface.launch(debug=True)
+    logger.debug("Launching Gradio interface")
+    interface.launch()