Update app.py

app.py

@@ -1,6 +1,45 @@
 import gradio as gr
 from huggingface_hub import InferenceClient
+import requests
 
+# Funktion, um Kontext von Wikipedia zu holen
+def get_wikipedia_summary(query):
+    try:
+        # Wir nutzen die öffentliche Wikipedia API
+        response = requests.get(
+            "https://de.wikipedia.org/w/api.php",
+            params={
+                "action": "query",
+                "format": "json",
+                "list": "search",
+                "srsearch": query,
+                "srlimit": 1
+            }
+        ).json()
+        
+        if not response["query"]["search"]:
+            return None
+            
+        page_id = response["query"]["search"][0]["pageid"]
+        
+        # Details zur Seite holen
+        details = requests.get(
+            "https://de.wikipedia.org/w/api.php",
+            params={
+                "action": "query",
+                "format": "json",
+                "prop": "extracts",
+                "pageids": page_id,
+                "explaintext": True,
+                "exintro": True,
+                "exsentences": 7 # Nur die ersten 7 Sätze
+            }
+        ).json()
+        
+        page = details["query"]["pages"][str(page_id)]
+        return page["extract"]
+    except Exception as e:
+        return None
 
 def respond(
     message,
@@ -10,61 +49,61 @@ def respond(
     temperature,
     top_p,
     hf_token: gr.OAuthToken,
+    use_wiki # Checkbox Input
 ):
-    """
-    For more information on `huggingface_hub` Inference API support, please check the docs: https://huggingface.co/docs/huggingface_hub/v0.22.2/en/guides/inference
-    """
-    client = InferenceClient(token=hf_token.token, model="openai/gpt-oss-20b")
+    client = InferenceClient(token=hf_token.token, model="meta-llama/Llama-3.2-1B-Instruct")
 
-    messages = [{"role": "system", "content": system_message}]
+    # --- HIER PASSIERT DAS IN-CONTEXT LEARNING ---
+    context_text = ""
+    if use_wiki:
+        wiki_content = get_wikipedia_summary(message)
+        if wiki_content:
+            context_text = f"\n\nEXTERNER KONTEXT (WIKIPEDIA): {wiki_content}\n"
+            gr.Info(f"Kontext gefunden: {wiki_content[:50]}...") # Kleines UI Feedback
+        else:
+            gr.Info("Kein Wikipedia-Artikel gefunden.")
 
-    messages.extend(history)
+    # Der Prompt zwingt das Modell, den Kontext zu nutzen
+    final_system_prompt = (
+        f"{system_message} "
+        f"Wenn 'EXTERNER KONTEXT' bereitgestellt wird, nutze dieses Wissen, um die Frage zu beantworten. "
+        f"Verlasse dich mehr auf den Kontext als auf dein eigenes Wissen."
+        f"{context_text}"
+    )
 
+    messages = [{"role": "system", "content": final_system_prompt}]
+    messages.extend(history)
     messages.append({"role": "user", "content": message})
 
     response = ""
-
-    for message in client.chat_completion(
-        messages,
-        max_tokens=max_tokens,
-        stream=True,
-        temperature=temperature,
-        top_p=top_p,
+    for msg in client.chat_completion(
+        messages, max_tokens=max_tokens, stream=True, temperature=temperature, top_p=top_p,
     ):
-        choices = message.choices
-        token = ""
-        if len(choices) and choices[0].delta.content:
-            token = choices[0].delta.content
-
-        response += token
-        yield response
-
-
-"""
-For information on how to customize the ChatInterface, peruse the gradio docs: https://www.gradio.app/docs/chatinterface
-"""
-chatbot = gr.ChatInterface(
-    respond,
-    type="messages",
-    additional_inputs=[
-        gr.Textbox(value="You are a friendly Chatbot.", label="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)",
-        ),
-    ],
-)
+        token = msg.choices[0].delta.content
+        if token:
+            response += token
+            yield response
 
+# --- GUI ---
 with gr.Blocks() as demo:
+    gr.Markdown("# 🧠 Der Wikipedia-gestützte Assistent")
+    gr.Markdown("Stelle eine Frage. Wenn du die Checkbox aktivierst, suche ich live nach Fakten!")
+    
+    with gr.Row():
+        wiki_checkbox = gr.Checkbox(label="Nutze Wikipedia-Wissen (RAG)", value=True)
+
+    chatbot = gr.ChatInterface(
+        respond,
+        additional_inputs=[
+            gr.Textbox(value="Du bist ein hilfreicher Assistent der Dinge genau und exakt erklärt.", label="System"),
+            gr.Slider(1, 1024, 512, label="Max Tokens"),
+            gr.Slider(0.1, 2.0, 0.7, label="Temp"),
+            gr.Slider(0.1, 1.0, 0.95, label="Top-p"),
+            wiki_checkbox
+        ]
+    )
     with gr.Sidebar():
         gr.LoginButton()
-    chatbot.render()
-
 
 if __name__ == "__main__":
-    demo.launch()
+    demo.launch()