import streamlit as st
import plotly.express as px
import pandas as pd
import streamlit_authenticator as stauth
import yaml
from yaml.loader import SafeLoader
import plotly.graph_objects as go
from transformers import pipeline
from PIL import Image, ImageDraw
from backend import *
# Start of Streamlit App
st.set_page_config(layout="centered")
hide_streamlit_style = '''
'''
st.markdown(hide_streamlit_style, unsafe_allow_html=True)
@st.cache_resource
# Function to initialise object_detection model
def initialise_object_detection_model():
checkpoint = "google/owlvit-base-patch32"
detector = pipeline(model=checkpoint, task="zero-shot-object-detection")
return detector
# Function to get result from object detection
def get_object_detection_results(detector, image_path, object_labels):
image = Image.open(image_path)
predictions = detector(
image,
candidate_labels=object_labels,
)
draw = ImageDraw.Draw(image)
for prediction in predictions:
box = prediction["box"]
label = prediction["label"]
score = prediction["score"]
xmin, ymin, xmax, ymax = box.values()
draw.rectangle((xmin, ymin, xmax, ymax), outline="red", width=1)
draw.text((xmin, ymin), f"{label}: {round(score,2)}", fill="white")
return image
detector = initialise_object_detection_model()
# Import configuration file for user authentication
with open('credentials.yaml') as file:
config = yaml.load(file, Loader=SafeLoader)
# Create an authentication object.
authenticator = stauth.Authenticate(
config['credentials'],
config['cookie']['name'],
config['cookie']['key'],
config['cookie']['expiry_days']
)
# List of advanced users
advanced_users = ['advanced']
# Landing page if user not logged in
if st.session_state['authentication_status'] is None:
# Landing page copy and banner
st.markdown('
Fly Situation Monitoring App 🪰
', unsafe_allow_html=True)
st.markdown('Keeping You Informed, Keeping Flies at Bay
', unsafe_allow_html=True)
st.write('\n')
st.write('\n')
# Loging log-in details
name, authentication_status, username = authenticator.login('', 'main')
# If log-in failed
if st.session_state['authentication_status'] is False:
st.error('Username/password is incorrect.')
st.write('\n')
st.write('\n')
# App if user is logged in and authenticated
if st.session_state['authentication_status']:
# Streamlit app start
st.title("Fly Situation Monitoring App 🪰")
st.markdown("Keeping You Informed, Keeping Flies at Bay")
st.write('\n')
# User selects a canteen
canteen = st.selectbox("Select a Canteen:", options=["Deck", "Frontier"])
st.write('\n')
# If user is a student, show basic app layout
if not st.session_state['username'] in advanced_users:
# Tabs
tab1, tab2, tab3 = st.tabs(["Current", "History", "FAQ"])
# Tab 1: Fly Situation
with tab1:
st.header("Current Fly Situation")
# Get data
fly_situation, delta1, delta2, delta3 = get_fly_situation(canteen)
# Display key information using cards
col1_fly_curr, col2_fly_curr, col3_fly_curr = st.columns(3)
col1_fly_curr.metric("Temperature", str(fly_situation["temperature"]) + " °C", delta=delta1)
col2_fly_curr.metric("Humidty", str(fly_situation["humidity"]) + " %", delta=delta2)
col3_fly_curr.metric("Fly Count", str(fly_situation["fly_count"]), delta=delta3, delta_color="inverse")
st.caption("Last updated at " + fly_situation["last_updated"] + " (5 min intervals)")
# Alert level
if fly_situation["fly_count"] > 20:
alert_level = "High 🔴"
alert_colour = "red"
elif fly_situation["fly_count"] > 10:
alert_level = "Moderate 🟠"
alert_colour = "orange"
else:
alert_level = "Low 🟢"
alert_colour = "green"
st.markdown(f"Alert Level: {alert_level}", unsafe_allow_html=True)
st.markdown('---')
# Camera locations
st.header("Smart Sensor Locations")
camera_locations = get_camera_locations(canteen)
st.map(camera_locations, size='size', zoom=18)
st.markdown('---')
# Feedback
st.header("Feedback")
# Gather feedback
feedback_col1, feedback_col2 = st.columns(2)
with feedback_col1:
user_feedback = st.text_area("Provide Feedback on the Fly Situation:")
with feedback_col2:
uploaded_files = st.file_uploader("Upload a Photo", accept_multiple_files=True, type=['jpg', 'png'])
for uploaded_file in uploaded_files:
st.image(uploaded_file)
if st.button("Submit Feedback"):
st.success("Feedback submitted successfully!")
st.write('\n')
st.write('\n')
st.write('\n')
# Tab 2: History
with tab2:
st.subheader("Fly Count Over Time")
# Get history data
fly_situation_history = get_fly_situation_history(canteen)
# Create a DataFrame for the time series data
df = pd.DataFrame(fly_situation_history)
sum_by_timestamp = df.groupby('timestamp')['fly_count'].sum().reset_index()
sum_by_timestamp["timestamp"] = pd.to_datetime(sum_by_timestamp["timestamp"])
# Plot the time series using Plotly Express
fig = px.line(sum_by_timestamp, x="timestamp", y="fly_count", labels={"fly_count": "Fly Count", "timestamp": "Timestamp"})
st.plotly_chart(fig)
# Question-and-Answer
with st.form("form"):
prompt = st.text_input("Ask a Question:")
submit = st.form_submit_button("Submit")
if prompt:
pass
#with st.spinner("Generating..."):
# Tab 3: FAQ
with tab3:
st.header("Frequently Asked Questions")
with st.expander("What is this app about?"):
st.write("This app provides you real-time information on fly activity by the smart fly monitoring system.")
with st.expander("How do the sensors work/detect fly activity?"):
st.write("The sensors built into the fly traps leverages cutting-edge AI methodologies for advanced fly detection.")
st.write("1) Object Detection - Using OWL-ViT, an open-vocabulary object detector, we can finetune the model specifically to recognise flies.")
st.write('\n')
st.write('\n')
st.write("Try OWL-ViT:")
object_labels = st.text_input("Enter your labels for the model to detect (comma-separated)", value="insect")
labels = object_labels.split(", ")
image_file = st.file_uploader("Upload an image", type=["jpg", "png"])
demo_image = st.checkbox("Load in demo image")
if image_file:
st.write('Before:')
st.image(image_file)
if image_file and object_labels:
st.write('After:')
with st.spinner("Detecting"):
st.image(image = get_object_detection_results(detector, image_file, labels))
if demo_image:
st.write('Before:')
st.image("images/fly.jpg")
if demo_image and object_labels:
st.write('After:')
with st.spinner("Detecting"):
st.image(image = get_object_detection_results(detector, "images/fly.jpg", labels))
st.write('\n')
st.write('\n')
st.write("2) Behaviour Analysis - By comparing consecutive frames, the system can extract data such as the trajectory, speed, and direction of each fly's movement. Training the system on these data can improve the system's detection of flies.")
trajectory_data = pd.DataFrame({
'X': [1, 2, 3, 4, 5],
'Y': [10, 25, 20, 25, 30],
'Timestamp': pd.date_range('2023-01-01', '2023-01-05', freq='D')
})
# Create a Plotly figure
fig = go.Figure()
# Add a trace for the trajectory
fig.add_trace(go.Scatter(x=trajectory_data['X'], y=trajectory_data['Y'], mode='lines'))
# Update layout
fig.update_layout(
xaxis_title='X-Coordinate',
yaxis_title='Y-Coordinate',
title='Example of a Fly Trajectory'
)
# Display the Plotly figure
st.plotly_chart(fig, use_container_width=True)
st.write('\n')
st.write('\n')
st.write('3) Training Augmentation - The fly detection system employs generative adversial networks, which generates synthetic fly images for training the fly detection model. This makes the system more robust at detecting flies in all scenarios.')
with st.expander("How accurate is the fly detection in the system?"):
st.write("The system is still in experimental phase.")
with st.expander("How often is the data updated or refreshed in real-time?"):
st.write("5 minute intervals.")
with st.expander("Why do I hear some sounds coming out from the fly traps?"):
st.write("The fly traps are built to emit accoustic sounds to attract flies.")
with st.expander("The traps seem to release some gas. What is that?"):
st.write("The fly traps release non-toxic pheremones that attract flies.")
# Logout
logout_col1, logout_col2 = st.columns([6,1])
with logout_col2:
st.write('\n')
st.write('\n')
st.write('\n')
authenticator.logout('Logout', 'main')
# Footer Credits
st.markdown('##')
st.markdown("---")
st.markdown("Created with ❤️ by HS2912 W4 Group 2")
else:
# Tabs
tab1, tab2, tab3 = st.tabs(["Current", "History", "Control System"])
# Tab 1: Fly Situation
with tab1:
st.header("Current Fly Situation")
# Get current data
fly_situation, delta1, delta2, delta3 = get_fly_situation(canteen)
# Display key information using cards
col1_fly_curr, col2_fly_curr, col3_fly_curr = st.columns(3)
col1_fly_curr.metric("Temperature", str(fly_situation["temperature"]) + " °C", delta=delta1)
col2_fly_curr.metric("Humidty", str(fly_situation["humidity"]) + " %", delta=delta2)
col3_fly_curr.metric("Fly Count", str(fly_situation["fly_count"]), delta=delta3, delta_color="inverse")
st.caption("Last updated at " + fly_situation["last_updated"] + " (5 min intervals)")
# Alert
if fly_situation["fly_count"] > 20:
alert_level = "High 🔴"
alert_colour = "red"
elif fly_situation["fly_count"] > 10:
alert_level = "Moderate 🟠"
alert_colour = "orange"
else:
alert_level = "Low 🟢"
alert_colour = "green"
st.markdown(f"Alert Level: {alert_level}", unsafe_allow_html=True)
st.markdown('---')
# Camera locations
st.header("Smart Sensor Locations")
camera_locations = get_camera_locations(canteen)
st.map(camera_locations, size='size', zoom=18)
st.markdown('---')
# Feedback
st.header("Feedback")
# Gather feedback
feedback_col1, feedback_col2 = st.columns(2)
with feedback_col1:
user_feedback = st.text_area("Provide Feedback on the Fly Situation:")
with feedback_col2:
uploaded_files = st.file_uploader("Upload a Photo", accept_multiple_files=True, type=['jpg', 'png'])
for uploaded_file in uploaded_files:
st.image(uploaded_file)
if st.button("Submit Feedback"):
st.success("Feedback submitted successfully!")
st.write('\n')
st.write('\n')
st.write('\n')
# Tab 2: History
with tab2:
# Fly count over time
st.subheader("Fly Count Over Time")
# Select sensor
selected_sensor = st.selectbox("Select Sensor:", ["All", "Sensor 1", "Sensor 2", "Sensor 3"])
# Get history data
fly_situation_history = get_fly_situation_history(canteen)
# Create a DataFrame for the time series data
df = pd.DataFrame(fly_situation_history)
if selected_sensor != "All":
df = df[df["sensor"]==int(selected_sensor[-1])]
sum_by_timestamp = df.groupby('timestamp')['fly_count'].sum().reset_index()
sum_by_timestamp["timestamp"] = pd.to_datetime(sum_by_timestamp["timestamp"])
# Plot the time series using Plotly Express
fig = px.line(sum_by_timestamp, x="timestamp", y="fly_count", labels={"fly_count": "Fly Count", "timestamp": "Timestamp"})
st.plotly_chart(fig)
# Pheremones level
st.subheader("Pheremone Level Over Time")
selected_sensor_level = st.selectbox("Select Sensor:", ["Sensor 1", "Sensor 2", "Sensor 3"])
# Get history data
sensor_pheremone_history = get_pheremone_levels(selected_sensor_level)
pheremone_df = pd.DataFrame(sensor_pheremone_history)
pheremone_df = pheremone_df[pheremone_df["sensor"] == int(selected_sensor_level[-1])]
pheremone_df['timestamp'] = pd.to_datetime(pheremone_df['timestamp'])
# Plot the time series using Plotly Express
fig = px.line(pheremone_df, x="timestamp", y="pheremone_level", labels={"pheremone_level": "Pheremone Level", "timestamp": "Timestamp"})
st.plotly_chart(fig)
# Question-and-Answer
with st.form("form"):
prompt = st.text_input("Ask a Question:")
submit = st.form_submit_button("Submit")
if prompt:
with st.spinner("Generating..."):
pass
# Tab 3: Control System
with tab3:
# Enable/disable automatic pest control system
st.header("System Settings")
automatic_control_enabled = st.toggle("Enable Automatic Pest Control", value=True)
st.write('\n')
st.write('\n')
if not automatic_control_enabled:
disabled = False
else:
disabled = True
# Camera
st.subheader("Smart Camera/Sensors")
sensor1 = st.toggle("Enable Sensor 1", value=True, disabled=disabled, key='deck_sensor_1')
sensor2 = st.toggle("Enable Sensor 2", value=True, disabled=disabled, key='deck_sensor_2')
sensor3 = st.toggle("Enable Sensor 3", value=True, disabled=disabled, key='deck_sensor_3')
st.write('\n')
# Audio
st.subheader("Audio")
# Accoustic
accoustic = st.selectbox("Accoustic Audio", ["Audio 1", "Audio 2", "Audio 3"], disabled=disabled)
st.write('\n')
# Pheremones
# Time interval for pheremones discharge in minutes)
st.subheader("Pheremones")
pheremones_interval = st.slider("Pheremones Discharge Interval (minutes)", min_value=5, max_value=60, value=15, step=5, disabled=disabled)
st.write('\n')
# Alerts
st.subheader('Alerts')
# Pest activity threshold for alerts
pest_activity_threshold = st.slider("Fly Count Threshold to Send Out Alerts", min_value=0, max_value=100, value=30, step=5, disabled=disabled)
st.write('\n')
# Instant alerts for pest sightings or unusual activity
st.markdown('Instant alert', unsafe_allow_html=True)
if st.button("Send Pest Alert", disabled=disabled):
st.success("Pest alert sent!")
st.write('\n')
# Notifications for upcoming preventive measures or scheduled treatments
st.markdown('Schedule notification for upcoming treatment day', unsafe_allow_html=True)
upcoming_event_date = st.date_input("Schedule Date", disabled=disabled)
upcoming_event_time = st.time_input("Set time for alert", disabled=disabled)
if st.button("Schedule Notification", disabled=disabled):
st.success(f"Notification scheduled for {upcoming_event_date} {upcoming_event_time}")
# Logout
logout_col1, logout_col2 = st.columns([6,1])
with logout_col2:
st.write('\n')
st.write('\n')
st.write('\n')
authenticator.logout('Logout', 'main')
# Footer Credits
st.markdown('##')
st.markdown("---")
st.markdown("Created with ❤️ by HS2912 W4 Group 2")