pages/6_Anomaly_Detection_Production.py

import streamlit as st
import json
import pandas as pd
import os
from utils import load_and_process_data, create_time_series_plot, display_statistics, call_api
import plotly.express as px
import plotly.graph_objects as go


if 'api_token' not in st.session_state:
    st.session_state.api_token =  os.getenv('NILM_API_TOKEN')

page_id = 6
if 'current_page' not in st.session_state:
    st.session_state.current_page = page_id
elif st.session_state.current_page != page_id:
    # Clear API response when switching to this page
    if 'api_response' in st.session_state:
        st.session_state.api_response = None
    # Update current page
    st.session_state.current_page = page_id

# Initialize session state variables
if 'current_file' not in st.session_state:
    st.session_state.current_file = None
if 'json_data' not in st.session_state:
    st.session_state.json_data = None
if 'api_response' not in st.session_state:
    st.session_state.api_response = None
if 'using_default_file' not in st.session_state:
    st.session_state.using_default_file = True

st.title("Energy Production Anomaly Detection")

st.markdown("""
This service analyzes energy production patterns to detect anomalies and unusual behavior in your data.

### Features
- Real-time anomaly detection
- Production irregularity identification
- Interactive visualization of detected anomalies
""")

# Default file path
default_file_path = "samples/6_anomaly_detection_production.json"  # Adjust this path to your default file

# File upload and processing
uploaded_file = st.file_uploader("Upload JSON file (or use default)", type=['json'])

# Load default file if no file is uploaded and using_default_file is True
if uploaded_file is None and st.session_state.using_default_file:
    if os.path.exists(default_file_path):
        st.info(f"Using default file: {default_file_path}")
        with open(default_file_path, 'r') as f:
            file_contents = f.read()
            if st.session_state.current_file != file_contents:
                st.session_state.current_file = file_contents
                st.session_state.json_data = json.loads(file_contents)
    else:
        st.warning(f"Default file not found at: {default_file_path}")
        st.session_state.using_default_file = False

# If a file is uploaded, process it
if uploaded_file:
    st.session_state.using_default_file = False
    try:
        file_contents = uploaded_file.read()
        st.session_state.current_file = file_contents
        st.session_state.json_data = json.loads(file_contents)
    except Exception as e:
        st.error(f"Error processing file: {str(e)}")

# Process and display data if available
if st.session_state.json_data:
    try:
        dfs = load_and_process_data(st.session_state.json_data)
        if dfs:
            st.header("Input Data Analysis")
            tabs = st.tabs(["Visualization", "Statistics", "Raw Data"])
            
            with tabs[0]:
                for unit, df in dfs.items():
                    st.plotly_chart(create_time_series_plot(df, unit), use_container_width=True)
                    
                    # Show basic statistical analysis
                    col1, col2, col3 = st.columns(3)
                    with col1:
                        st.metric("Average Production", 
                                f"{df['datacellar:value'].mean():.2f} {unit}")
                    with col2:
                        st.metric("Standard Deviation", 
                                f"{df['datacellar:value'].std():.2f} {unit}")
                    with col3:
                        st.metric("Total Samples", 
                                len(df))
            
            with tabs[1]:
                display_statistics(dfs)
            
            with tabs[2]:
                st.json(st.session_state.json_data)
            
            # Add analysis options
            st.subheader("Anomaly Detection")
            col1, col2 = st.columns(2)
            with col1:
                if st.button("Detect Anomalies", key="detect_button"):
                    if not st.session_state.api_token:
                        st.error("Please enter your API token in the sidebar first.")
                    else:
                        with st.spinner("Analyzing production patterns..."):
                            # Add sensitivity and window_size to the request
                            modified_data = st.session_state.json_data.copy()
                           
                            # Convert back to JSON and call API
                            modified_content = json.dumps(modified_data).encode('utf-8')
                            st.session_state.api_response = call_api(
                                modified_content,
                                st.session_state.api_token,
                                "inference_production_ad"
                            )

    except Exception as e:
        st.error(f"Error processing data: {str(e)}")

# Display API results
if st.session_state.api_response:
    st.header("Anomaly Detection Results")
    tabs = st.tabs(["Anomaly Visualization", "Raw Results"])
    
    with tabs[0]:
        response_dfs = load_and_process_data(
            st.session_state.api_response,
            input_data=st.session_state.json_data
        )
        if response_dfs:
            anomalies = response_dfs['boolean']
            anomalies = anomalies[anomalies['datacellar:value']==True]
            
            del response_dfs['boolean']
            for unit, df in response_dfs.items():
                fig = create_time_series_plot(df, unit, service_type="Anomaly Detection")
                # Get df values for anomalies
                anomaly_df = df.iloc[anomalies['datacellar:timeStamp'].index]
                
                fig.add_trace(go.Scatter(
                    x=anomaly_df['datacellar:timeStamp'], 
                    y=anomaly_df['datacellar:value'], 
                    mode='markers', 
                    marker=dict(color='red'), 
                    name='Anomalies'
                ))
                
                # Create visualization with highlighted anomalies
                st.plotly_chart(
                    fig,
                    use_container_width=True
                )
    
    with tabs[1]:
        st.json(st.session_state.api_response)