streamlit.py

import streamlit as st # web development
import numpy as np # np mean, np random 
import pandas as pd # read csv, df manipulation
import time # to simulate a real time data, time loop 
import plotly.express as px # interactive charts 
import paho.mqtt.client as mqtt
import json
import warnings
from tensorflow.keras.models import load_model
import joblib
import plotly.graph_objects as go
warnings.filterwarnings('ignore')


model = load_model(r"src\rtu\models\lstm_4rtu_smooth_02.keras")
scaler = joblib.load(r'src\rtu\models\scaler_1.pkl')
# kmeans = joblib.load('kmeans_model.pkl')
kmeans1 = joblib.load(r'src\rtu\models\kmeans_model1.pkl')
kmeans2 = joblib.load(r'src\rtu\models\kmeans_model2.pkl')
kmeans3 = joblib.load(r'src\rtu\models\kmeans_model3.pkl')
kmeans4 = joblib.load(r'src\rtu\models\kmeans_model4.pkl')
# pca = joblib.load('pca_model.pkl')

st.set_page_config(
    page_title = 'Real-Time Data Buliding 59',
    page_icon = '✅',
    layout = 'wide'
)

st.title("Buliding 59 Dashboard")
placeholder = st.empty()

broker_address = "localhost"
broker_port = 1883
topic = "sensor_data"

df = pd.DataFrame(columns=['hp_hws_temp',
 'rtu_003_sa_temp',
 'rtu_003_oadmpr_pct',
 'rtu_003_ra_temp',
 'rtu_003_oa_temp',
 'rtu_003_ma_temp',
 'rtu_003_sf_vfd_spd_fbk_tn',
 'rtu_003_rf_vfd_spd_fbk_tn',
 'rtu_004_sa_temp',
 'rtu_004_oadmpr_pct',
 'rtu_004_ra_temp',
 'rtu_004_oa_temp',
 'rtu_004_ma_temp',
 'rtu_004_sf_vfd_spd_fbk_tn',
 'rtu_004_rf_vfd_spd_fbk_tn',
 'rtu_001_sa_temp',
 'rtu_001_oadmpr_pct',
 'rtu_001_ra_temp',
 'rtu_001_oa_temp',
 'rtu_001_ma_temp',
 'rtu_001_sf_vfd_spd_fbk_tn',
 'rtu_001_rf_vfd_spd_fbk_tn',
 'rtu_002_sa_temp',
 'rtu_002_oadmpr_pct',
 'rtu_002_ra_temp',
 'rtu_002_oa_temp',
 'rtu_002_ma_temp',
 'rtu_002_sf_vfd_spd_fbk_tn',
 'rtu_002_rf_vfd_spd_fbk_tn',
#  'rtu_004_sat_sp_tn',
#  'rtu_003_sat_sp_tn',
#  'rtu_001_sat_sp_tn',
#  'rtu_002_sat_sp_tn',
 'air_temp_set_1',
 'air_temp_set_2',
 'dew_point_temperature_set_1d',
 'relative_humidity_set_1',
 'solar_radiation_set_1'])
actual_list = [0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0]
pred_list = [0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0]
resid_list = [0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0]
distance = [0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0]
pca_x = [0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0]
pca_y = [0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0]

def on_message(client, userdata, message):
    global df
    payload = json.loads(message.payload.decode())
    
    hp_hws_temp = payload['hp_hws_temp']
    rtu_003_sa_temp = payload['rtu_003_sa_temp']
    rtu_003_oadmpr_pct = payload['rtu_003_oadmpr_pct']
    rtu_003_ra_temp = payload['rtu_003_ra_temp']
    rtu_003_oa_temp = payload['rtu_003_oa_temp']
    rtu_003_ma_temp = payload['rtu_003_ma_temp']
    rtu_003_sf_vfd_spd_fbk_tn = payload['rtu_003_sf_vfd_spd_fbk_tn']
    rtu_003_rf_vfd_spd_fbk_tn =payload['rtu_003_rf_vfd_spd_fbk_tn']
    rtu_004_sa_temp = payload['rtu_004_sa_temp']
    rtu_004_oadmpr_pct = payload['rtu_004_oadmpr_pct']
    rtu_004_ra_temp = payload['rtu_004_ra_temp']
    rtu_004_oa_temp = payload['rtu_004_oa_temp']
    rtu_004_ma_temp = payload['rtu_004_ma_temp']
    rtu_004_sf_vfd_spd_fbk_tn = payload['rtu_004_sf_vfd_spd_fbk_tn']
    rtu_004_rf_vfd_spd_fbk_tn = payload['rtu_004_rf_vfd_spd_fbk_tn']
    rtu_001_sa_temp = payload['rtu_001_sa_temp']
    rtu_001_oadmpr_pct = payload['rtu_001_oadmpr_pct']
    rtu_001_ra_temp = payload['rtu_001_ra_temp']
    rtu_001_oa_temp = payload['rtu_001_oa_temp']
    rtu_001_ma_temp = payload['rtu_001_ma_temp']
    rtu_001_sf_vfd_spd_fbk_tn = payload['rtu_001_sf_vfd_spd_fbk_tn']
    rtu_001_rf_vfd_spd_fbk_tn =payload['rtu_001_rf_vfd_spd_fbk_tn']
    rtu_002_sa_temp = payload['rtu_002_sa_temp']
    rtu_002_oadmpr_pct = payload['rtu_002_oadmpr_pct']
    rtu_002_ra_temp = payload['rtu_002_ra_temp']
    rtu_002_oa_temp = payload['rtu_002_oa_temp']
    rtu_002_ma_temp = payload['rtu_002_ma_temp']
    rtu_002_sf_vfd_spd_fbk_tn = payload['rtu_002_sf_vfd_spd_fbk_tn']
    rtu_002_rf_vfd_spd_fbk_tn = payload['rtu_002_rf_vfd_spd_fbk_tn']
    # rtu_004_sat_sp_tn = payload['rtu_004_sat_sp_tn']
    # rtu_003_sat_sp_tn = payload['rtu_003_sat_sp_tn']
    # rtu_001_sat_sp_tn = payload['rtu_001_sat_sp_tn']
    # rtu_002_sat_sp_tn = payload['rtu_002_sat_sp_tn']
    air_temp_set_1 = payload['air_temp_set_1']
    air_temp_set_2 = payload['air_temp_set_2']
    dew_point_temperature_set_1d = payload['dew_point_temperature_set_1d']
    relative_humidity_set_1 = payload['relative_humidity_set_1']
    solar_radiation_set_1 = payload['solar_radiation_set_1']
    
    len_df = len(df)
    df.loc[len_df] = {'hp_hws_temp':hp_hws_temp,
    'rtu_003_sa_temp':rtu_003_sa_temp,
    'rtu_003_oadmpr_pct': rtu_003_oadmpr_pct,
    'rtu_003_ra_temp':rtu_003_ra_temp,
    'rtu_003_oa_temp': rtu_003_oa_temp,
    'rtu_003_ma_temp': rtu_003_ma_temp,
    'rtu_003_sf_vfd_spd_fbk_tn': rtu_003_sf_vfd_spd_fbk_tn,
    'rtu_003_rf_vfd_spd_fbk_tn':rtu_003_rf_vfd_spd_fbk_tn,
    'rtu_004_sa_temp':rtu_004_sa_temp,
    'rtu_004_oadmpr_pct':rtu_004_oadmpr_pct,
    'rtu_004_ra_temp':rtu_004_ra_temp,
    'rtu_004_oa_temp':rtu_004_oa_temp,
    'rtu_004_ma_temp':rtu_004_ma_temp,
    'rtu_004_sf_vfd_spd_fbk_tn':rtu_004_sf_vfd_spd_fbk_tn,
    'rtu_004_rf_vfd_spd_fbk_tn':rtu_004_rf_vfd_spd_fbk_tn,
    'rtu_001_sa_temp':rtu_001_sa_temp,
    'rtu_001_oadmpr_pct': rtu_001_oadmpr_pct,
    'rtu_001_ra_temp':rtu_001_ra_temp,
    'rtu_001_oa_temp': rtu_001_oa_temp,
    'rtu_001_ma_temp': rtu_001_ma_temp,
    'rtu_001_sf_vfd_spd_fbk_tn': rtu_001_sf_vfd_spd_fbk_tn,
    'rtu_001_rf_vfd_spd_fbk_tn':rtu_001_rf_vfd_spd_fbk_tn,
    'rtu_002_sa_temp':rtu_002_sa_temp,
    'rtu_002_oadmpr_pct':rtu_002_oadmpr_pct,
    'rtu_002_ra_temp':rtu_002_ra_temp,
    'rtu_002_oa_temp':rtu_002_oa_temp,
    'rtu_002_ma_temp':rtu_002_ma_temp,
    'rtu_002_sf_vfd_spd_fbk_tn':rtu_002_sf_vfd_spd_fbk_tn,
    'rtu_002_rf_vfd_spd_fbk_tn':rtu_002_rf_vfd_spd_fbk_tn,
    # 'rtu_004_sat_sp_tn':rtu_004_sat_sp_tn,
    # 'rtu_003_sat_sp_tn' :rtu_003_sat_sp_tn,
    # 'rtu_001_sat_sp_tn':rtu_001_sat_sp_tn,
    # 'rtu_002_sat_sp_tn':rtu_002_sat_sp_tn,
    'air_temp_set_1':air_temp_set_1,
    'air_temp_set_2':air_temp_set_2,
    'dew_point_temperature_set_1d':dew_point_temperature_set_1d,
    'relative_humidity_set_1':relative_humidity_set_1,
    'solar_radiation_set_1':solar_radiation_set_1}
    
    if len_df>30:
        df_window = df[len_df-31:len_df]
        df_window = df_window.astype('float32')
        df_trans = scaler.transform(df_window)
        df_new = df_trans[:30,:].reshape((1,30,34))#
        pred = model.predict(df_new)
        pred_copy = pred.copy()
        actual = df_trans[30,:29]#
        resid = actual - pred
        #---------
        pred.resize((pred.shape[0], pred.shape[1] + len(df_trans[30,29:])))#
        pred[:, -len(df_trans[30,29:]):] = df_trans[30,29:]#
        pred = scaler.inverse_transform(np.array(pred))
        actual = scaler.inverse_transform(np.array([df_trans[30,:]]))
        #---------
        actual_list.pop(0)
        pred_list.pop(0)
        resid_list.pop(0)
        # distance.pop(0)
        # pca_x.pop(0)
        # pca_y.pop(0)
        actual_list.append(actual[0,1])
        pred_list.append(pred[0,1])
        resid_list.append(resid[0,1])
        # distance.append(np.linalg.norm(pred_copy-kmeans.cluster_centers_[0], ord=2, axis = 1))
        # dist_color = [1 if num >= 5 else 0 for num in distance]
        dist = []
        dist.append(np.linalg.norm(resid[:,1:8]-kmeans1.cluster_centers_[0], ord=2, axis = 1))
        dist.append(np.linalg.norm(resid[:,8:15]-kmeans2.cluster_centers_[0], ord=2, axis = 1))
        dist.append(np.linalg.norm(resid[:,15:22]-kmeans3.cluster_centers_[0], ord=2, axis = 1))
        dist.append(np.linalg.norm(resid[:,22:29]-kmeans4.cluster_centers_[0], ord=2, axis = 1))
        dist = np.array(dist)
        # dist_color = [1 if num >= 2 else 0 for num in dist]
        
        # pca_cord = pca.transform(resid) 
        # pca_x.append(pca_cord[0,0])
        # pca_y.append(pca_cord[0,1])
        # clust_center = pca.transform(kmeans.cluster_centers_)
        # theta = np.linspace(0, 2*np.pi, 100)
        # radius = 2
        # x_circle = clust_center[0, 0] + radius * np.cos(theta)
        # y_circle = clust_center[0, 1] + radius * np.sin(theta)
        
        ind = np.linspace(1, 30, 30)
        
        with placeholder.container():
            col1, fig_col1, fig_col2 = st.columns(3)
            with col1:
                st.header("RTU Status")
                for i in range(4):
                    rtu = ['RTU 1', 'RTU 2', 'RTU 3', 'RTU 4']
                    tol = [2,2,2,0.1]#[4.5,4,5,5]
                    status_icon = "🔧" if dist[i,0] > tol[i] else "🔄"
                    status = "Damper or Fan issue" if dist[i,0] > tol[i] else "Normal"
                    status_markdown = f"**{rtu[i]} {status_icon}**\n\nSA Temp: {int(actual[0,1])}°C\nRA Temp: {int(actual[0,3])}°C\n\nStatus: {status}"
                    st.markdown(status_markdown, unsafe_allow_html=True)
            with fig_col1:
                # st.markdown("### Fault")
                st.header("Fault")
                fig1 = go.Figure()
                # fig1.add_trace(go.Scatter(x=ind, y=resid_list, mode='lines', name='Actual',line=dict(color='blue')))
                # fig1 = px.scatter(x=pca_x, y=pca_y,color=dist_color,color_discrete_map={'1': 'red', '0': 'green'})
                # fig1.add_trace(go.Scatter(x=x_circle, y=y_circle, mode='lines',line=dict(color='green')))
                colors = ['red' if value > 0.1 else 'blue' for value in dist[:, 0]]
                fig1 = fig1.add_trace(go.Bar(x=['RTU 1', 'RTU 2', 'RTU 3', 'RTU 4'],y=dist[:, 0],marker_color=colors,width=0.4))
                fig1.update_layout(width=500,height=400 )
                st.write(fig1)
            with fig_col2:
                st.header("Mixed Air temperature")
                fig2 = go.Figure()
                fig2.add_trace(go.Scatter(x=ind, y=actual_list, mode='lines', name='Actual',line=dict(color='blue')))
                fig2.add_trace(go.Scatter(x=ind, y=pred_list, mode='lines', name='Predicted',line=dict(color='red', dash='dot')))
                fig2.update_layout(yaxis_range=[50, 80],width=500,height=400 )
                st.write(fig2)
            
            st.markdown("### Detailed Data View")
            st.dataframe(df[len_df-5:len_df])
            # time.sleep(1)
 


client = mqtt.Client(mqtt.CallbackAPIVersion.VERSION1)
client.on_message = on_message
client.connect(broker_address, broker_port)
client.subscribe(topic)
client.loop_forever()