import streamlit as st
import time
from core.farnsworth import Farnsworth
st.set_page_config(page_title="Farnsworth AI", layout="wide", page_icon="⚡")
# Custom CSS for "Premium Aesthetics"
st.markdown("""
<style>
.stApp {
background: linear-gradient(to right, #0f2027, #203a43, #2c5364);
color: #ffffff;
}
.stChatMessage {
background-color: rgba(255, 255, 255, 0.05);
border-radius: 10px;
padding: 10px;
}
h1 {
font-family: 'Helvetica Neue', sans-serif;
text-align: center;
background: -webkit-linear-gradient(45deg, #00d2ff, #3a7bd5);
-webkit-background-clip: text;
-webkit-text-fill-color: transparent;
}
</style>
""", unsafe_allow_html=True)
@st.cache_resource
def get_farnsworth():
return Farnsworth()
ai = get_farnsworth()
# Sidebar: System Status
with st.sidebar:
st.header("🧠 Neural Status")
# Memory Stats
mem_count = ai.memory.index.ntotal if hasattr(ai.memory, 'index') else 0
st.metric("LTM Vectors", mem_count)
node_count = ai.memory.graph.number_of_nodes()
edge_count = ai.memory.graph.number_of_edges()
st.metric("Knowledge Nodes", node_count)
st.metric("Synaptic Connections", edge_count)
st.divider()
st.header("📥 Ingestion")
url = st.text_input("Ingest URL")
if st.button("Absorb Knowledge"):
if url:
status = ai.ingest_data(url)
st.success(status)
if st.button("Ingest Local Data"):
status = ai.ingest_data("local")
st.success(status)
# Main Chat
st.title("Farnsworth AI v1.0")
st.caption("Autonomous • Recursive • Evolving")
if "messages" not in st.session_state:
st.session_state.messages = []
for msg in st.session_state.messages:
with st.chat_message(msg["role"]):
st.markdown(msg["content"])
if prompt := st.chat_input("Command the swarm..."):
st.session_state.messages.append({"role": "user", "content": prompt})
with st.chat_message("user"):
st.markdown(prompt)
with st.chat_message("assistant"):
message_placeholder = st.empty()
full_response = ""
# Stream response (simulated for now, as Swarm runs atomically)
with st.spinner("Swarm Processing..."):
response = ai.chat(prompt)
# Typewriter effect
for chunk in response.split():
full_response += chunk + " "
time.sleep(0.02)
message_placeholder.markdown(full_response + "▌")
message_placeholder.markdown(full_response)
st.session_state.messages.append({"role": "assistant", "content": full_response})
