app.py

import streamlit as st
import json
import ee
import os
import pandas as pd
import geopandas as gpd
from datetime import datetime, timedelta
import leafmap.foliumap as leafmap
import time
import re

# Set up the page layout
st.set_page_config(layout="wide")

# Custom button styling
m = st.markdown(
    """
    <style>
    div.stButton > button:first-child {
        background-color: #006400;
        color:#ffffff;
    }
    </style>""",
    unsafe_allow_html=True,
)

# Logo
st.write(
    f"""
    <div style="display: flex; justify-content: space-between; align-items: center;">
        <img src="https://huggingface.co/spaces/YashMK89/GEE_Calculator/resolve/main/ISRO_Logo.png"  style="width: 20%; margin-right: auto;">
        <img src="https://huggingface.co/spaces/YashMK89/GEE_Calculator/resolve/main/SAC_Logo.png"  style="width: 20%; margin-left: auto;">
    </div>
    """,
    unsafe_allow_html=True,
)

# Authenticate and initialize Earth Engine
earthengine_credentials = os.environ.get("EE_Authentication")

# Initialize Earth Engine with secret credentials
os.makedirs(os.path.expanduser("~/.config/earthengine/"), exist_ok=True)
with open(os.path.expanduser("~/.config/earthengine/credentials"), "w") as f:
    f.write(earthengine_credentials)

ee.Initialize(project='ee-yashsacisro24')

# Load Sentinel dataset options from JSON file
with open("sentinel_datasets.json") as f:
    data = json.load(f)

# Display the title and dataset selection
st.title("Sentinel Dataset")

# Select dataset category and subcategory
main_selection = st.selectbox("Select Sentinel Dataset Category", list(data.keys()))

if main_selection:
    sub_options = data[main_selection]["sub_options"]
    sub_selection = st.selectbox("Select Specific Dataset ID", list(sub_options.keys()))

# Earth Engine Index Calculator Section
st.header("Earth Engine Index Calculator")

# Choose Index or Custom Formula
index_choice = st.selectbox("Select an Index or Enter Custom Formula", ['NDVI', 'NDWI', 'Average NO₂', 'Custom Formula'])

# Initialize custom_formula variable
custom_formula = ""

# Display corresponding formula based on the index selected
if index_choice == 'NDVI':
    st.write("Formula for NDVI: NDVI = (B8 - B4) / (B8 + B4)")
elif index_choice == 'NDWI':
    st.write("Formula for NDWI: NDWI = (B3 - B8) / (B3 + B8)")
elif index_choice == 'Average NO₂':
    st.write("Formula for Average NO₂: Average NO₂ = Mean(NO2 band)")
elif index_choice == 'Custom Formula':
    custom_formula = st.text_input("Enter Custom Formula (e.g., 'B5 - B4 / B5 + B4')")
    st.write(f"Custom Formula: {custom_formula}")  # Display the custom formula after the user inputs it

# Function to read points from CSV
def read_csv(file_path):
    df = pd.read_csv(file_path)
    return df

# Function to read points from GeoJSON
def read_geojson(file_path):
    gdf = gpd.read_file(file_path)
    return gdf

# Function to read points from KML
def read_kml(file_path):
    gdf = gpd.read_file(file_path, driver='KML')
    return gdf

# Ask user whether they want to process 'Point' or 'Polygon' data
shape_type = st.selectbox("Do you want to process 'Point' or 'Polygon' data?", ["Point", "Polygon"])

# Ask user to upload a file based on shape type
file_upload = st.file_uploader(f"Upload your {shape_type} data (CSV, GeoJSON, KML)", type=["csv", "geojson", "kml"])

# Date Input for Start and End Dates
start_date = st.date_input("Start Date", value=pd.to_datetime('2020-01-01'))
end_date = st.date_input("End Date", value=pd.to_datetime('2020-12-31'))

# Convert start_date and end_date to string format for Earth Engine
start_date_str = start_date.strftime('%Y-%m-%d')
end_date_str = end_date.strftime('%Y-%m-%d')

# Frequency selection
frequency = st.selectbox("Select Frequency for Date Incrementation", 
                         ["Daily", "Weekly", "Monthly", "Yearly", "Custom Number of Days"])

# Custom days input if "Custom Number of Days" is selected
if frequency == "Custom Number of Days":
    custom_days = st.number_input("Enter number of days", min_value=1, value=1)
else:
    custom_days = None  # Default to None when custom days aren't required

# Initialize session state for storing results if not already done
if 'results' not in st.session_state:
    st.session_state.results = []
if 'last_params' not in st.session_state:
    st.session_state.last_params = {}
if 'map_data' not in st.session_state:
    st.session_state.map_data = None  # Initialize map_data

# Function to check if parameters have changed
def parameters_changed():
    return (
        st.session_state.last_params.get('main_selection') != main_selection or
        st.session_state.last_params.get('sub_selection') != sub_selection or
        st.session_state.last_params.get('index_choice') != index_choice or
        st.session_state.last_params.get('start_date_str') != start_date_str or
        st.session_state.last_params.get('end_date_str') != end_date_str
    )

# If parameters have changed, reset the results
if parameters_changed():
    st.session_state.results = []  # Clear the previous results
    # Update the last parameters to the current ones
    st.session_state.last_params = {
        'main_selection': main_selection,
        'sub_selection': sub_selection,
        'index_choice': index_choice,
        'start_date_str': start_date_str,
        'end_date_str': end_date_str
    }

# Function to perform index calculations
def calculate_ndvi(image, geometry):
    ndvi = image.normalizedDifference(['B8', 'B4']).rename('NDVI')
    result = ndvi.reduceRegion(
        reducer=ee.Reducer.mean(),
        geometry=geometry,
        scale=30
    )
    return result.get('NDVI')

def calculate_ndwi(image, geometry):
    ndwi = image.normalizedDifference(['B3', 'B8']).rename('NDWI')
    result = ndwi.reduceRegion(
        reducer=ee.Reducer.mean(),
        geometry=geometry,
        scale=30
    )
    return result.get('NDWI')

def calculate_avg_no2_sentinel5p(image, geometry):
    no2 = image.select('NO2').reduceRegion(
        reducer=ee.Reducer.mean(),
        geometry=geometry,
        scale=1000
    ).get('NO2')
    return no2

def calculate_custom_formula(image, geometry, formula):
    result = image.expression(formula).rename('Custom Index').reduceRegion(
        reducer=ee.Reducer.mean(),
        geometry=geometry,
        scale=30
    )
    return result.get('Custom Index')

# Process each point for index calculation
if file_upload:
    locations_df = None  # Initialize locations_df to None
    polygons_df = None   # Initialize polygons_df to None

    file_extension = os.path.splitext(file_upload.name)[1].lower()

    # Read file based on shape type
    if shape_type == 'Point':
        if file_extension == '.csv':
            locations_df = read_csv(file_upload)
        elif file_extension == '.geojson':
            locations_df = read_geojson(file_upload)
        elif file_extension == '.kml':
            locations_df = read_kml(file_upload)
        else:
            st.error("Unsupported file type. Please upload a CSV, GeoJSON, or KML file for points.")
    elif shape_type == 'Polygon':
        if file_extension == '.geojson':
            polygons_df = read_geojson(file_upload)
        elif file_extension == '.kml':
            polygons_df = read_kml(file_upload)
        else:
            st.error("Unsupported file type. Please upload a GeoJSON or KML file for polygons.")

    if locations_df is not None:
        # Display a preview of the points data
        st.write("Preview of the uploaded points data:")
        st.dataframe(locations_df.head())

        # Create a LeafMap object to display the points
        m = leafmap.Map(center=[locations_df['latitude'].mean(), locations_df['longitude'].mean()], zoom=10)

        # Add points to the map using a loop
        for _, row in locations_df.iterrows():
            latitude = row['latitude']
            longitude = row['longitude']
            
            # Check if latitude or longitude are NaN and skip if they are
            if pd.isna(latitude) or pd.isna(longitude):
                continue  # Skip this row and move to the next one
            
            m.add_marker(location=[latitude, longitude], popup=row.get('name', 'No Name'))

        # Display map
        st.write("Map of Uploaded Points:")
        m.to_streamlit()

        # Store the map in session_state
        st.session_state.map_data = m

            
    # Loop through the selected frequency and increment the date
    while start_date <= end_date:
        start_date_str = start_date.strftime('%Y-%m-%d')
        
    # Store results with dates for aggregation
    results_with_dates = []
    for idx, row in locations_df.iterrows():
        latitude = row['latitude']
        longitude = row['longitude']
        location_name = row.get('name', f"Location_{idx}")

        # Skip processing if latitude or longitude is NaN
        if pd.isna(latitude) or pd.isna(longitude):
            continue  # Skip this row and move to the next one

        # Define the region of interest (ROI)
        roi = ee.Geometry.Point([longitude, latitude])

        # Load Sentinel-2 image collection
        collection = ee.ImageCollection(sub_options[sub_selection]) \
            .filterDate(ee.Date(start_date_str), ee.Date(end_date_str)) \
            .filterBounds(roi)

        # Check if the collection has images for the selected date range
        image_count = collection.size().getInfo()
        if image_count == 0:
            st.warning(f"No images found for {location_name}.")
        else:
            st.write(f"Found {image_count} images for {location_name}.")
            image = collection.first()

            # Perform the calculation based on user selection
            result = None
            if index_choice == 'NDVI':
                result = calculate_ndvi(image, roi)
            elif index_choice == 'NDWI':
                result = calculate_ndwi(image, roi)
            elif index_choice == 'Average NO₂':
                if 'NO2' in image.bandNames().getInfo():
                    result = calculate_avg_no2_sentinel5p(image, roi)
                else:
                    st.warning(f"No NO2 band found for {location_name}. Please use Sentinel-5P for NO₂ data.")
            elif index_choice == 'Custom Formula' and custom_formula:
                result = calculate_custom_formula(image, roi, custom_formula)

            if result is not None:
                calculated_value = result.getInfo()  # Get the numeric value

                # Add the date and calculated value to the results list
                st.session_state.results.append({
                    'Location Name': location_name,
                    'Latitude': latitude,
                    'Longitude': longitude,
                    'Date': date,
                    'Calculated Value': calculated_value
                })
        
        # Increment the current date based on the selected frequency
        if frequency == "Daily":
            current_date += timedelta(days=1)
        elif frequency == "Weekly":
            current_date += timedelta(weeks=1)
        elif frequency == "Monthly":
            current_date += timedelta(days=30)  # Approximation
        elif frequency == "Yearly":
            current_date += timedelta(days=365)  # Approximation
        elif frequency == "Custom Number of Days" and custom_days is not None:
            current_date += timedelta(days=custom_days)

# After processing, show the results
if st.session_state.results:
    # Convert the results to a DataFrame for better visualization
    result_df = pd.DataFrame(st.session_state.results)

    # Show the results in a table format
    st.write("Processed Results Table:")
    st.dataframe(result_df[['Location Name', 'Latitude', 'Longitude', 'Date', 'Calculated Value']])

    # Generate the dynamic filename
    filename = f"{main_selection}_{sub_selection}_{start_date.strftime('%Y/%m/%d')}_{end_date.strftime('%Y/%m/%d')}_{shape_type}.csv"

    # Convert results to DataFrame for download
    if st.session_state.results:
        result_df = pd.DataFrame(st.session_state.results)
        st.download_button(
            label="Download results as CSV",
            data=result_df.to_csv(index=False).encode('utf-8'),
            file_name=filename,
            mime='text/csv'
        )