Update app.py

app.py

@@ -1,8 +1,6 @@
 """
-CodeNyx – HF-Space-ready chatbot
-- 3 B parameters
-- 4-bit quant → < 8 GB VRAM
-- Accepts HF token only for gated models (none required here)
+CodeNyx – bullet-proof HF-Space chatbot
+Public model → no token required by default
 """
 
 import os
@@ -17,26 +15,25 @@ from transformers import (
 from huggingface_hub import login
 from threading import Thread
 
-# ------------------------------------------------------------------
-# 1. Configuration
-# ------------------------------------------------------------------
-BOT_NAME      = "CodeNyx"
-MODEL_ID      = "bigcode/starcoder2-3b"        #  <- public, no token needed
+BOT_NAME = "CodeNyx"
+MODEL_ID = "bigcode/starcoder2-3b"           # PUBLIC
+MAX_NEW_TOK = 1024
+TEMPERATURE = 0.2
+TOP_P = 0.9
+
 SYSTEM_PROMPT = (
     f"You are {BOT_NAME}, an expert open-source coding assistant. "
     "Always provide concise, runnable code snippets with short explanations."
 )
-MAX_NEW_TOK   = 1024
-TEMPERATURE   = 0.2
-TOP_P         = 0.9
 
 # ------------------------------------------------------------------
-# 2. Lightweight 4-bit loader
+# 1. Model loader with None-safe token
 # ------------------------------------------------------------------
-def load_model(token: str = None):
-    """Return (tokenizer, model) or raise."""
+def load_model(token: str | None):
+    """Return (tok, model) or raise."""
+    token = (token or "").strip() or None          # None-safe
     if token:
-        login(token)  # only if provided
+        login(token)
 
     bnb_config = BitsAndBytesConfig(
         load_in_4bit=True,
@@ -45,35 +42,35 @@ def load_model(token: str = None):
         bnb_4bit_use_double_quant=True,
     )
 
-    tokenizer = AutoTokenizer.from_pretrained(
+    tok = AutoTokenizer.from_pretrained(
         MODEL_ID,
-        use_auth_token=token or None,
+        use_auth_token=token,
         trust_remote_code=True,
     )
-    if tokenizer.pad_token is None:
-        tokenizer.pad_token = tokenizer.eos_token
+    if tok.pad_token is None:
+        tok.pad_token = tok.eos_token
 
-    model = AutoModelForCausalLM.from_pretrained(
+    mdl = AutoModelForCausalLM.from_pretrained(
         MODEL_ID,
         quantization_config=bnb_config,
         device_map="auto",
-        use_auth_token=token or None,
+        use_auth_token=token,
         trust_remote_code=True,
     )
-    return tokenizer, model
+    return tok, mdl
 
 # ------------------------------------------------------------------
-# 3. Chat helpers
+# 2. Chat helpers
 # ------------------------------------------------------------------
+tokenizer, model = None, None
+
 def build_prompt(history, user_input):
     messages = [{"role": "system", "content": SYSTEM_PROMPT}]
     for human, ai in history:
-        messages.append({"role": "user", "content": human})
-        messages.append({"role": "assistant", "content": ai})
+        messages += [{"role": "user", "content": human},
+                     {"role": "assistant", "content": ai}]
     messages.append({"role": "user", "content": user_input})
-    return tokenizer.apply_chat_template(
-        messages, tokenize=False, add_generation_prompt=True
-    )
+    return tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
 
 def user_turn(user_message, history):
     return "", history + [[user_message, ""]]
@@ -82,12 +79,7 @@ def bot_turn(history):
     prompt = build_prompt(history[:-1], history[-1][0])
     inputs = tokenizer(prompt, return_tensors="pt").to(model.device)
 
-    streamer = TextIteratorStreamer(
-        tokenizer,
-        skip_prompt=True,
-        skip_special_tokens=True,
-    )
-
+    streamer = TextIteratorStreamer(tokenizer, skip_prompt=True, skip_special_tokens=True)
     gen_kwargs = dict(
         **inputs,
         max_new_tokens=MAX_NEW_TOK,
@@ -106,27 +98,16 @@ def bot_turn(history):
         yield history
 
 # ------------------------------------------------------------------
-# 4. Gradio UI
+# 3. Gradio UI
 # ------------------------------------------------------------------
-tokenizer, model = None, None  # lazy load
-
 with gr.Blocks(title=f"{BOT_NAME} – AI Pair-Programmer") as demo:
-    gr.Markdown(f"""
-    # 🤖 {BOT_NAME} – AI Pair-Programmer  
-    *Public model — no token needed.*  
-    Ask any coding question and get **runnable code + short explanations**.
-    """)
-
-    # Optional token box (for future gated models)
-    token_box = gr.Textbox(label="🤗 HF Token (optional)", type="password", visible=False)
-    load_btn  = gr.Button("Load / Reload", visible=False)
-    status_lbl = gr.Label(value="Loading model …")
+    gr.Markdown(f"# 🤖 {BOT_NAME} – AI Pair-Programmer\n*3 B params, public model, zero-config*")
 
-    # Chat area
-    chatbot = gr.Chatbot(height=450)
+    status_lbl = gr.Label(value="Loading model …")
+    chatbot = gr.Chatbot(height=500)
     with gr.Row():
         msg = gr.Textbox(
-            placeholder="Ask me to write, debug, or explain code …",
+            placeholder="Ask me to write / debug / explain code …",
             lines=2,
             scale=8,
             show_label=False,
@@ -135,21 +116,15 @@ with gr.Blocks(title=f"{BOT_NAME} – AI Pair-Programmer") as demo:
         send_btn = gr.Button("Send", scale=1, variant="primary")
     clear_btn = gr.Button("🗑️ Clear")
 
-    # ------------------------------------------------------------------
-    # 5. Events
-    # ------------------------------------------------------------------
-    def _load(token):
+    def _load():
         global tokenizer, model
         try:
-            tokenizer, model = load_model(token if token.strip() else None)
+            tokenizer, model = load_model(None)   # None = no token
             return "✅ Model loaded!"
         except Exception as e:
             return f"❌ {e}"
 
-    load_btn.click(_load, token_box, status_lbl)
-
-    # auto-load once on start
-    demo.load(_load, token_box, status_lbl)
+    demo.load(_load, None, status_lbl)
 
     def _send(user_msg, hist):
         return user_turn(user_msg, hist)
@@ -162,7 +137,7 @@ with gr.Blocks(title=f"{BOT_NAME} – AI Pair-Programmer") as demo:
     clear_btn.click(lambda: None, None, chatbot)
 
 # ------------------------------------------------------------------
-# 6. Launch
+# 4. Launch
 # ------------------------------------------------------------------
 if __name__ == "__main__":
     demo.queue().launch(server_name="0.0.0.0", server_port=7860, share=True)