Create app.py

app.py

@@ -0,0 +1,71 @@
+import streamlit as st
+from huggingface_hub import InferenceClient
+
+# Constants
+SPACE_URL = "https://z7svds7k42bwhhgm.us-east-1.aws.endpoints.huggingface.cloud"
+HF_API_KEY = HF_API_KEY = os.getenv("HF_API_KEY")
+DUBS_PATH = "🐾"  # Optional: Replace with an avatar path if needed
+
+# Streamlit Configuration
+st.set_page_config(page_title="Chatbot Test", page_icon="🤖", layout="centered")
+
+# Chat State
+if "messages" not in st.session_state:
+    st.session_state["messages"] = []
+
+# Function to Stream Response
+def stream_response(prompt_text, api_key):
+    """
+    Stream text from the HF Inference Endpoint using the InferenceClient.
+    Yields each partial chunk of text as it arrives.
+    """
+    client = InferenceClient(SPACE_URL, token=api_key)
+
+    gen_kwargs = {
+        "max_new_tokens": 512,
+        "top_k": 30,
+        "top_p": 0.9,
+        "temperature": 0.2,
+        "repetition_penalty": 1.02,
+        "stop_sequences": ["<|endoftext|>"]
+    }
+
+    stream = client.text_generation(prompt_text, stream=True, details=True, **gen_kwargs)
+
+    try:
+        for response in stream:
+            if response.token.special:
+                continue
+            yield response.token.text
+    except Exception as e:
+        yield f"Error: {e}"
+
+# Streamlit Chat Interface
+st.title("Chatbot Testing Interface")
+
+# User Input Section
+prompt = st.chat_input("Enter your message...")
+
+if prompt:
+    # 1) Add the user's message to session state
+    st.session_state["messages"].append({"role": "user", "content": prompt})
+    st.chat_message("user").write(prompt)
+
+    # 2) Build combined chat history for the model prompt
+    chat_history = "".join(
+        [f"<|{msg['role']}|>{msg['content']}<|end|>" for msg in st.session_state["messages"]]
+    )
+
+    # 3) Generate the response
+    with st.chat_message("assistant", avatar=DUBS_PATH):
+        with st.spinner("Dubs is thinking... Woof Woof! 🐾"):
+            full_response = ""
+            placeholder = st.empty()  # Placeholder for streaming response
+            response = stream_response(chat_history, HF_API_KEY)
+            for item in response:
+                full_response += item
+                placeholder.markdown(full_response)
+            placeholder.markdown(full_response)
+
+    # 4) Add assistant response to the session state
+    st.session_state["messages"].append({"role": "assistant", "content": full_response})