Update app.py

app.py

@@ -2,119 +2,75 @@ import gradio as gr
 from huggingface_hub import InferenceClient
 import time
 
-# Cliente da Inference API
 client = InferenceClient("lambdaindie/lambdai")
 
-# Função para responder no chatbot
-def respond(
-    message,
-    history: list[tuple[str, str]],
-    system_message,
-    max_tokens,
-    temperature,
-    top_p,
-):
-    messages = [{"role": "system", "content": system_message}]
+def respond(message, history, system_message, max_tokens, temperature, top_p):
+    # Build base message history
+    messages = [{"role": "system", "content": system_message}] if system_message else []
 
-    # Adicionando a história da conversa
-    for val in history:
-        if val[0]:
-            messages.append({"role": "user", "content": val[0]})
-        if val[1]:
-            messages.append({"role": "assistant", "content": val[1]})
+    for user, assistant in history:
+        if user:
+            messages.append({"role": "user", "content": user})
+        if assistant:
+            messages.append({"role": "assistant", "content": assistant})
 
-    messages.append({"role": "user", "content": message})
+    # Phase 1 — Thinking aloud (reasoning step)
+    thinking_prompt = messages + [
+        {
+            "role": "user",
+            "content": f"{message}\n\nThink step-by-step before answering."
+        }
+    ]
 
-    response = ""
+    reasoning = ""
+    yield "**Thinking...**\n```markdown\n```"  # Trigger gray markdown block
 
-    # Mostrando o indicador de "thinking"
-    yield "Thinking... 🤔"
-    time.sleep(1)  # Atraso artificial para simular tempo de processamento
-
-    # Fluxo de resposta do cliente da API
-    for message in client.chat_completion(
-        messages,
+    for chunk in client.chat_completion(
+        thinking_prompt,
         max_tokens=max_tokens,
         stream=True,
         temperature=temperature,
         top_p=top_p,
     ):
-        token = message.choices[0].delta.content
-        response += token
-        yield response
+        token = chunk.choices[0].delta.content or ""
+        reasoning += token
+        yield f"**Thinking...**\n```markdown\n{reasoning.strip()}```"
+
+    time.sleep(0.5)  # Optional dramatic pause
 
+    # Phase 2 — Final answer
+    final_prompt = messages + [
+        {"role": "user", "content": message},
+        {"role": "assistant", "content": reasoning.strip()},
+        {"role": "user", "content": "Now answer based on your reasoning above."}
+    ]
+
+    final_answer = ""
+    for chunk in client.chat_completion(
+        final_prompt,
+        max_tokens=max_tokens,
+        stream=True,
+        temperature=temperature,
+        top_p=top_p,
+    ):
+        token = chunk.choices[0].delta.content or ""
+        final_answer += token
+        yield final_answer.strip()
 
-# Interface do Gradio com chat customizado
 demo = gr.ChatInterface(
     respond,
+    title="LENIRΛ",
+    theme=gr.themes.Base(primary_hue="gray", font=["JetBrains Mono", "monospace"]),
     additional_inputs=[
-        gr.Textbox(value="", label="System message", lines=1, placeholder="System message..."),
-        gr.Slider(minimum=1, maximum=2048, value=512, step=1, label="Max new tokens"),
-        gr.Slider(minimum=0.1, maximum=4.0, value=0.7, step=0.1, label="Temperature"),
-        gr.Slider(
-            minimum=0.1,
-            maximum=1.0,
-            value=0.95,
-            step=0.05,
-            label="Top-p (nucleus sampling)",
+        gr.Textbox(
+            value="You are a concise, logical AI that explains its reasoning clearly before answering.",
+            label="System Message"
         ),
-    ],
-    title="Lambda-v1-1B",  # Título na interface
-    description="Chatbot alimentado pelo modelo Lambdai",  # Descrição simples
-    theme="dark",  # Usando o tema Dark do Gradio
+        gr.Slider(64, 2048, value=512, step=1, label="Max Tokens"),
+        gr.Slider(0.1, 2.0, value=0.7, step=0.1, label="Temperature"),
+        gr.Slider(0.1, 1.0, value=0.95, step=0.05, label="Top-p")
+    ]
 )
 
-# Customização de CSS simples para JetBrains Mono e visual Dark
-demo.css = """
-    * {
-        font-family: 'JetBrains Mono', monospace;
-    }
-    .gradio-container {
-        background-color: #121212;
-        color: #ffffff;
-        border-radius: 10px;
-        padding: 20px;
-    }
-    .chatbox {
-        background-color: #181818;
-        border-radius: 8px;
-        color: #f5f5f5;
-        border: 1px solid #333;
-    }
-    .gr-button {
-        background-color: #4a90e2;
-        color: white;
-        border-radius: 5px;
-        padding: 10px 20px;
-        font-size: 16px;
-    }
-    .gr-button:hover {
-        background-color: #357ab7;
-    }
-    .gr-slider {
-        background-color: #333;
-        color: #f5f5f5;
-        border-radius: 8px;
-    }
-    .gr-chatbox-container {
-        background-color: #1f1f1f;
-        border-radius: 10px;
-    }
-    .gr-output {
-        font-family: 'JetBrains Mono', monospace;
-        color: #f5f5f5;
-    }
-    .gr-input {
-        font-family: 'JetBrains Mono', monospace;
-        color: #f5f5f5;
-    }
-    .gr-chatbox .message {
-        font-family: 'JetBrains Mono', monospace;
-    }
-    .gr-button.gr-loading {
-        background-color: #f39c12;
-    }
-"""
-
 if __name__ == "__main__":
     demo.launch()