Update app.py

app.py

@@ -1,14 +1,14 @@
 import os
 import math
-import requests
-from typing import List, Dict, Any, Tuple
+from typing import List, Dict, Any, Tuple, Optional
 
+import requests
 import gradio as gr
 from openai import OpenAI
 
 # -------------------- CONFIG --------------------
 
-CHAT_MODEL = "gpt-5.1"  # Change here if your OpenAI model ID differs
+CHAT_MODEL = "gpt-5"  # main chat model
 EMBED_MODEL = "text-embedding-3-large"
 
 DEFAULT_SYSTEM_PROMPT = """You are a Retrieval-Augmented Generation (RAG) assistant.
@@ -17,7 +17,7 @@ Rules:
 - Answer ONLY using the provided knowledge base context and system instructions.
 - If the answer is not clearly supported by the context, say "I don’t know based on the current knowledge base."
 - Do not invent sources, statistics, or facts that are not present in the context.
-- When applicable, cite which source you used (e.g., "According to the uploaded PDF" or "Based on zenai.world").
+- When applicable, cite which source you used (e.g., "According to the uploaded file" or "Based on zenai.world").
 - Be clear, concise, and structured.
 """
 
@@ -29,15 +29,18 @@ PRESET_CONFIGS = {
     },
     "ZEN Sites Deep QA (zenai.world + AI Arena)": {
         "system": DEFAULT_SYSTEM_PROMPT
-        + "\n\nYou specialize in answering questions about ZEN AI’s mission, programs, and AI Arena.",
+        + "\n\nYou specialize in answering questions about ZEN AI’s mission, programs, AI Pioneer, and ZEN AI Arena.",
         "urls": "https://zenai.world\nhttps://us.zenai.biz",
-        "text": "ZEN AI builds the first global AI × Web3 literacy movement with youth, homeschool, and professional tracks.",
+        "text": (
+            "ZEN AI is building the first global AI × Web3 literacy and automation movement, "
+            "with youth, homeschool, and professional tracks and blockchain-verified credentials."
+        ),
     },
     "Policy Explainer (external PDFs / links)": {
         "system": DEFAULT_SYSTEM_PROMPT
-        + "\n\nYou act as a neutral policy explainer. Summarize clearly, highlight key risks and opportunities.",
+        + "\n\nYou act as a neutral policy explainer. Summarize clearly, highlight key risks, opportunities, and practical implications.",
         "urls": "",
-        "text": "This preset is for uploading AI policy PDFs, legal texts, and reports.",
+        "text": "This preset is for uploading AI policy PDFs, legal texts, and governance reports.",
     },
     "Research Notebook / Personal RAG Sandbox": {
         "system": DEFAULT_SYSTEM_PROMPT
@@ -47,8 +50,8 @@ PRESET_CONFIGS = {
     },
 }
 
+# -------------------- TEXT / EMBEDDING HELPERS --------------------
 
-# -------------------- HELPER FUNCTIONS --------------------
 
 def chunk_text(text: str, max_chars: int = 2000, overlap: int = 200) -> List[str]:
     """Simple character-based chunking with overlap."""
@@ -74,25 +77,27 @@ def cosine_similarity(a: List[float], b: List[float]) -> float:
         return 0.0
     dot = sum(x * y for x, y in zip(a, b))
     norm_a = math.sqrt(sum(x * x for x in a))
-    norm_b = math.sqrt(sum(y * y for y in b))
+    norm_b = math.sqrt(sum(x * x for x in b))
     if norm_a == 0 or norm_b == 0:
         return 0.0
     return dot / (norm_a * norm_b)
 
 
+# -------------------- DATA SOURCE HELPERS --------------------
+
+
 def fetch_url_text(url: str) -> str:
-    """Fetch text from a URL in a very lightweight way."""
+    """Fetch text from a URL in a lightweight way."""
     try:
-        resp = requests.get(url, timeout=10)
+        resp = requests.get(url, timeout=12)
         resp.raise_for_status()
-        # crude HTML stripping: keep text only
         text = resp.text
-        # Remove basic tags
+
+        # crude HTML stripping: cut off at first script/style and remove angle brackets
         for tag in ["<script", "<style"]:
             if tag in text:
-                # Truncate at first occurrence of script/style to avoid junk
                 text = text.split(tag)[0]
-        # Replace angle brackets
+
         text = text.replace("<", " ").replace(">", " ")
         return text
     except Exception as e:
@@ -108,13 +113,14 @@ def read_file_text(path: str) -> str:
         if any(path_lower.endswith(ext) for ext in [".txt", ".md", ".csv", ".json"]):
             with open(path, "r", encoding="utf-8", errors="ignore") as f:
                 return f.read()
-        # If you want to support PDFs or DOCX, you can add optional parsing here,
-        # but we avoid extra dependencies to keep the app robust.
         return f"[Unsupported file type for RAG content: {os.path.basename(path)}]"
     except Exception as e:
         return f"[Error reading file {os.path.basename(path)}: {e}]"
 
 
+# -------------------- EMBEDDING / KB BUILD --------------------
+
+
 def build_embeddings(
     api_key: str,
     docs: List[Dict[str, Any]],
@@ -124,13 +130,14 @@ def build_embeddings(
         return [], "⚠️ No documents to index."
 
     client = OpenAI(api_key=api_key)
-    kb_chunks = []
+    kb_chunks: List[Dict[str, Any]] = []
     total_chunks = 0
 
     for d in docs:
         source = d.get("source", "unknown")
         text = d.get("text", "")
         chunks = chunk_text(text, max_chars=2000, overlap=200)
+
         for idx, ch in enumerate(chunks):
             try:
                 emb_resp = client.embeddings.create(
@@ -148,7 +155,6 @@ def build_embeddings(
                 )
                 total_chunks += 1
             except Exception as e:
-                # Keep going even if one embedding fails
                 kb_chunks.append(
                     {
                         "id": f"{source}_{idx}_error",
@@ -183,7 +189,7 @@ def retrieve_context(
     except Exception as e:
         return "", f"⚠️ Error creating query embedding: {e}"
 
-    scored = []
+    scored: List[Tuple[float, Dict[str, Any]]] = []
     for d in kb:
         emb = d.get("embedding") or []
         if not emb:
@@ -215,6 +221,7 @@ def retrieve_context(
 
 # -------------------- GRADIO CALLBACKS --------------------
 
+
 def save_api_key(api_key: str):
     api_key = (api_key or "").strip()
     if not api_key:
@@ -233,13 +240,13 @@ def build_knowledge_base(
     api_key: str,
     urls_text: str,
     raw_text: str,
-    file_paths: List[str] | None,
+    file_paths: Optional[List[str]],
 ):
     api_key = (api_key or "").strip()
     if not api_key:
         return "❌ Please save your OpenAI API key first.", []
 
-    docs = []
+    docs: List[Dict[str, Any]] = []
 
     # URLs
     urls = [u.strip() for u in (urls_text or "").splitlines() if u.strip()]
@@ -252,9 +259,7 @@ def build_knowledge_base(
         docs.append({"source": "Raw Text Block", "text": raw_text})
 
     # Files
-    if file_paths is not None:
-        if isinstance(file_paths, str):
-            file_paths = [file_paths]
+    if file_paths:
         for p in file_paths:
             if not p:
                 continue
@@ -295,7 +300,8 @@ def chat_with_rag(
     client = OpenAI(api_key=api_key)
 
     # Assemble messages for OpenAI
-    messages = []
+    messages: List[Dict[str, str]] = []
+
     combined_system = (
         DEFAULT_SYSTEM_PROMPT.strip()
         + "\n\n---\n\nUser System Instructions:\n"
@@ -312,7 +318,7 @@ def chat_with_rag(
         )
         messages.append({"role": "system", "content": context_block})
 
-    # Add truncated history for conversational continuity
+    # Add truncated history
     recent_history = history[-10:] if history else []
     for msg in recent_history:
         if msg.get("role") in ("user", "assistant"):
@@ -332,7 +338,6 @@ def chat_with_rag(
     except Exception as e:
         answer = f"⚠️ OpenAI API error: {e}"
 
-    # Update history for display and next turn
     new_history = history + [
         {"role": "user", "content": user_message},
         {"role": "assistant", "content": answer},
@@ -347,10 +352,10 @@ def clear_chat():
 
 # -------------------- UI LAYOUT --------------------
 
-with gr.Blocks(title="RAG Chatbot — GPT-5.1 + URLs / Files / Text") as demo:
+with gr.Blocks(title="RAG Chatbot — GPT-5 + URLs / Files / Text") as demo:
     gr.Markdown(
         """
-# 🔍 RAG Chatbot — GPT-5.1 + URLs / Files / Text
+# 🔍 RAG Chatbot — GPT-5 + URLs / Files / Text
 
 1. Enter your **OpenAI API key** and click **Save**.  
 2. Add knowledge via **URLs**, **uploaded files**, and/or **raw text**.  
@@ -417,7 +422,7 @@ with gr.Blocks(title="RAG Chatbot — GPT-5.1 + URLs / Files / Text") as demo:
             gr.Markdown("### 💬 RAG Chat")
 
             chatbot = gr.Chatbot(
-                label="RAG Chatbot (GPT-5.1)",
+                label="RAG Chatbot (GPT-5)",
                 type="messages",
                 height=450,
             )
@@ -457,13 +462,14 @@ with gr.Blocks(title="RAG Chatbot — GPT-5.1 + URLs / Files / Text") as demo:
         outputs=[kb_status_md, kb_state],
     )
 
-    # Wiring: chat send
+    # Wiring: chat send (button)
     send_btn.click(
         fn=chat_with_rag,
         inputs=[user_input, api_key_state, kb_state, system_box, chat_state],
         outputs=[chatbot, chat_state, debug_md],
     )
 
+    # Wiring: chat send (Enter key)
     user_input.submit(
         fn=chat_with_rag,
         inputs=[user_input, api_key_state, kb_state, system_box, chat_state],