Update app.py

app.py

@@ -1,64 +1,66 @@
-import gradio as gr
-from huggingface_hub import InferenceClient
-
-"""
-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("HuggingFaceH4/zephyr-7b-beta")
-
 
-def respond(
-    message,
-    history: list[tuple[str, str]],
-    system_message,
-    max_tokens,
-    temperature,
-    top_p,
-):
-    messages = [{"role": "system", "content": system_message}]
+import os
+import requests
+from bs4 import BeautifulSoup
+import gradio as gr
+from langchain_cohere import ChatCohere
+from langchain_core.messages import HumanMessage, SystemMessage
+from langfuse import Langfuse
+from langfuse.callback import CallbackHandler
+from langchain.callbacks.manager import CallbackManager
 
-    for val in history:
-        if val[0]:
-            messages.append({"role": "user", "content": val[0]})
-        if val[1]:
-            messages.append({"role": "assistant", "content": val[1]})
+# ENV VARS must be set in Hugging Face Space settings for API key and Langfuse
+api_key = os.environ.get("COHERE_API_KEY")
+langfuse_public_key = os.environ.get("LANGFUSE_PUBLIC_KEY")
+langfuse_secret_key = os.environ.get("LANGFUSE_SECRET_KEY")
 
-    messages.append({"role": "user", "content": message})
+os.environ["LANGFUSE_PUBLIC_KEY"] = langfuse_public_key
+os.environ["LANGFUSE_SECRET_KEY"] = langfuse_secret_key
+os.environ["LANGFUSE_HOST"] = "https://cloud.langfuse.com"
 
-    response = ""
+# Load policy content from Cohere Docs
+def fetch_policy():
+    url = "https://docs.cohere.com/docs/cohere-labs-acceptable-use-policy"
+    try:
+        response = requests.get(url, timeout=10)
+        response.raise_for_status()
+        soup = BeautifulSoup(response.text, "html.parser")
+        content_div = soup.find("div", class_="markdown") or soup.find("main") or soup
+        return content_div.get_text(separator="\n").strip()[:1500]
+    except Exception as e:
+        return "⚠️ Failed to retrieve policy. " + str(e)
 
-    for message in client.chat_completion(
-        messages,
-        max_tokens=max_tokens,
-        stream=True,
-        temperature=temperature,
-        top_p=top_p,
-    ):
-        token = message.choices[0].delta.content
+context = fetch_policy()
 
-        response += token
-        yield response
+# Setup Langfuse and LLM
+langfuse_callback = CallbackHandler()
+llm = ChatCohere(
+    cohere_api_key=api_key,
+    model="command-r-plus",
+    callback_manager=CallbackManager([langfuse_callback]),
+    verbose=True,
+)
+langfuse = Langfuse()
 
+def run_policy_check(query: str):
+    trace = langfuse.trace(name="RAG_CommandRPlus_Trace", user_id="huggingface-demo")
+    messages = [
+        SystemMessage(content=f"This is the policy context from Cohere:\n{context}"),
+        HumanMessage(content=query),
+    ]
+    response = llm.invoke(messages)
+    trace.score(name="response_quality", value=1)
+    trace.update(status="completed")
+    return response.content
 
-"""
-For information on how to customize the ChatInterface, peruse the gradio docs: https://www.gradio.app/docs/chatinterface
-"""
-demo = gr.ChatInterface(
-    respond,
-    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)",
-        ),
-    ],
+# Gradio interface
+demo = gr.Interface(
+    fn=run_policy_check,
+    inputs=gr.Textbox(label="Ask a question about acceptable use of Command R+", lines=3),
+    outputs=gr.Textbox(label="Cohere Command R+ Answer"),
+    title="Policy QA with Cohere Command R+",
+    description="This agent retrieves the Cohere Acceptable Use Policy and answers your ethical use-case questions using LangChain + Command R+.",
 )
 
-
 if __name__ == "__main__":
-    demo.launch()
+    demo.launch()