Update app.py

app.py

@@ -1,21 +1,26 @@
 import os, requests, gradio as gr
 from huggingface_hub import InferenceClient
 
+# Using a high-capacity model for better tabular summaries
 client = InferenceClient("meta-llama/Llama-3.3-70B-Instruct", token=os.getenv("HF_TOKEN"))
 
 def safe_int(val, default=0):
+    """Data normalization helper."""
     try:
         return int(val) if val is not None else default
     except (ValueError, TypeError):
         return default
 
 def fetch_catalog_detailed(query, limit):
-    """Fetches and normalizes data based on user limit."""
+    """
+    Retrieves exactly the number of results requested by the user.
+    """
     try:
-        # Use the limit provided by the slider
+        # Crucial: Passing the 'limit' directly to the API
         url = f"https://openlibrary.org/search.json?q={query}&limit={limit}"
-        resp = requests.get(url, timeout=5).json()
-        docs = resp.get("docs", [])
+        response = requests.get(url, timeout=5)
+        response.raise_for_status()
+        docs = response.json().get("docs", [])
         
         books = []
         for d in docs:
@@ -26,85 +31,107 @@ def fetch_catalog_detailed(query, limit):
                 "editions": safe_int(d.get("edition_count"), 0)
             })
         return books
-    except:
+    except Exception as e:
+        print(f"API Error: {e}")
         return []
 
-# --- FUNCTION ARGUMENTS MUST MATCH gr.submit(inputs=[...]) ORDER ---
 def librarian_agent(message, history, session_state, sort_by, num_results):
+    """
+    Main Agent Logic. 
+    Order of arguments MUST match the 'inputs' list in msg.submit below.
+    """
+    if not message:
+        return history, session_state
+
     if session_state is None:
         session_state = {"verified": False, "age": None}
 
-    # 1. Gate Logic
+    # --- 1. ETHICAL GATE ---
     if not session_state["verified"]:
         age_digits = "".join(filter(str.isdigit, message))
         if age_digits:
             session_state["age"] = int(age_digits)
             session_state["verified"] = True
-            reply = "✅ Verified. What topic are we researching today?"
+            reply = "✅ **Verified.** My archives are now open. What subject can I help you with?"
         else:
-            reply = "👋 Please enter your age to start."
+            reply = "👋 Welcome. To comply with safety guidelines, please enter your **age** to begin."
         history.append({"role": "assistant", "content": reply})
         return history, session_state
 
-    # 2. Retrieval with the Slider Value
+    # --- 2. RAG RETRIEVAL (Respecting Slider) ---
     raw_books = fetch_catalog_detailed(message, int(num_results))
     
-    # 3. Sorting Logic
+    # --- 3. DETERMINISTIC SORTING ---
     if sort_by == "Newest First":
         raw_books = sorted(raw_books, key=lambda x: x['year'], reverse=True)
     elif sort_by == "Popularity":
         raw_books = sorted(raw_books, key=lambda x: x['editions'], reverse=True)
 
-    # 4. Context Preparation
+    # Prepare data for LLM
     catalog_summary = ""
     for b in raw_books:
-        year_str = b['year'] if b['year'] > 1000 else "Unknown"
-        catalog_summary += f"BOOK: {b['title']} | BY: {b['author']} | YEAR: {year_str} | EDITIONS: {b['editions']}\n"
+        yr = b['year'] if b['year'] > 1000 else "Unknown"
+        catalog_summary += f"BOOK: {b['title']} | AUTHOR: {b['author']} | YEAR: {yr} | EDITIONS: {b['editions']}\n"
 
-    # 5. LLM Synthesis
+    # --- 4. AGENTIC SYNTHESIS ---
     llm_messages = [
         {
             "role": "system", 
             "content": (
-                f"You are a librarian for a {session_state['age']} year old. "
-                "Output ONLY a Markdown table: | Book & Author | Year | Editions | Summary |."
+                f"You are a professional librarian for a {session_state['age']}-year-old. "
+                "You provide high-density information. "
+                "Respond ONLY with a Markdown table: | Book & Author | Year | Editions | Summary |."
+                "Do not mention internal lists. Summaries must be 1-2 lines of insight."
             )
         },
-        {"role": "user", "content": f"Catalog:\n{catalog_summary}\n\nUser: {message}"}
+        {"role": "user", "content": f"INTERNAL CATALOG DATA:\n{catalog_summary}\n\nUSER QUERY: {message}"}
     ]
 
     try:
         output = client.chat_completion(messages=llm_messages, max_tokens=1500)
         bot_res = output.choices[0].message.content
     except Exception as e:
-        bot_res = f"Service error: {str(e)}"
+        bot_res = f"The library is experiencing technical issues: {str(e)}"
 
     history.append({"role": "user", "content": message})
     history.append({"role": "assistant", "content": bot_res})
     return history, session_state
 
-# --- GRADIO UI ---
-with gr.Blocks() as demo:
-    gr.Markdown("## 📚 AI Librarian Agent")
+# --- GRADIO INTERFACE ---
+with gr.Blocks(title="AI Librarian Agent") as demo:
+    gr.Markdown("## 📚 AI Librarian Agent (RAG Optimized)")
     state = gr.State(None)
     
     with gr.Row():
         # Input 4
-        sort_option = gr.Dropdown(["Relevance", "Newest First", "Popularity"], label="Sort", value="Relevance")
+        sort_option = gr.Dropdown(
+            choices=["Relevance", "Newest First", "Popularity"], 
+            label="Sort Priority", 
+            value="Relevance"
+        )
         # Input 5
-        result_count = gr.Slider(minimum=3, maximum=15, step=1, value=5, label="Count")
+        result_count = gr.Slider(
+            minimum=3, 
+            maximum=15, 
+            step=1, 
+            value=5, 
+            label="Results to Retrieve"
+        )
     
     # Input 2
-    chatbot = gr.Chatbot(label="Catalog", value=[{"role": "assistant", "content": "👋 Enter your **age**."}])
+    chatbot = gr.Chatbot(
+        label="Librarian Consultation", 
+        value=[{"role": "assistant", "content": "👋 Please enter your **age** to access the library."}]
+    )
     # Input 1
-    msg = gr.Textbox(label="Query", placeholder="Search...")
+    msg = gr.Textbox(label="Message", placeholder="Enter age first, then search books...")
 
-    # THE CRITICAL STEP: Mapping UI components to function arguments
-    # 1: msg -> message
-    # 2: chatbot -> history
-    # 3: state -> session_state
-    # 4: sort_option -> sort_by
-    # 5: result_count -> num_results
+    # INPUT MAPPING CHECK:
+    # 1. message (msg)
+    # 2. history (chatbot)
+    # 3. session_state (state)
+    # 4. sort_by (sort_option)
+    # 5. num_results (result_count)
     msg.submit(
         fn=librarian_agent, 
         inputs=[msg, chatbot, state, sort_option, result_count],