data_utils.py

import pandas as pd
import numpy as np
import geopandas as gpd

import requests

url = "https://Projects-by-IF-model-temp-api-v2.hf.space/model-effect"
headers = {"Content-Type": "application/json"}

cell_area = 100*100 #in square meters

def make_XYT(df, treatment_col = 'Pct_CoberturaVeg', target_col = 'LST'):
    
    Y = df[target_col].values

    T = df[treatment_col].values

    X = df.drop(columns=[treatment_col, target_col]).values
    
    
    return X, Y, T
    
def prepare_data(og_file):
    original_df = pd.read_csv(og_file, index_col=0)
    # df = pd.read_csv(training_file, index_col=0)

    original_df = original_df.round(2)
    
    df = original_df[['Elevacion', 'Neighbor_NDBI', 'Proximidad_agua',
       'Neighbor_%CoberturaVeg', 'Pct_CoberturaVeg', 'LST']].copy()
    df = df.astype(np.float32)

    X, Y, T = make_XYT(df)
    
    return original_df, df, X, Y, T


def prepare_treatment_df(original_df, df):
    treatment_df = original_df[['x','y','Pct_CoberturaVeg']].copy()
    treatment_df.loc[:, 'Pct_Construccion'] = original_df['Pct_Construccion'].values
    treatment_df.loc[:, 'LST'] = df['LST'].values
    
    treatment_df = treatment_df.reset_index()
    return treatment_df
    

def run_treatment_increase(X, T, original_df, df, value, name, simulation_mode):
    
    treatment_df = prepare_treatment_df(original_df, df)

    if simulation_mode == "Usar control deslizante":
        T_sim = T.copy()  
        value_num = 1+ (value/100)
        T_sim = T_sim*value_num
    else:
        T_sim = original_df['simulacion_arboles']
    data = {"X": X.tolist(), "T0":T.tolist(), "T1":(T_sim).tolist()}
    response = requests.post(url, headers=headers, json=data)
    if response.ok:
        result = response.json()
        # print("Model response:", result["effect"])
    else:
        print("Request failed:", response.status_code, response.text)
    
    treatment_df[name] = result["effect"]
    treatment_df['pct_total_trees'] = T_sim
    treatment_df['pp_trees_increase'] = treatment_df['pct_total_trees'] - treatment_df['Pct_CoberturaVeg']

    return treatment_df

def process_data_trees_to_temp(value, simulation_mode, original_df, df, X, Y, T):
    name = 'temp_decrease'
    treatment_df = run_treatment_increase(X, T, original_df, df, value, name, simulation_mode)

    treatment_df['new_temp'] = treatment_df['LST'] + treatment_df[name]

    #add columns translated to square meters
    treatment_df['sqm_trees_increase'] = (treatment_df['pp_trees_increase']/100)*cell_area
    treatment_df['sqm_total_trees'] = (treatment_df['pct_total_trees']/100)*cell_area

    col_order = ['index', 'x', 'y', 'Pct_CoberturaVeg', 'Pct_Construccion', 'LST','temp_decrease', 
    'new_temp', 'pp_trees_increase','pct_total_trees', 'sqm_trees_increase','sqm_total_trees',]  # to match gdf trees

    treatment_df = treatment_df[col_order]

    gdf = gpd.GeoDataFrame(treatment_df, geometry=gpd.points_from_xy(treatment_df.x, treatment_df.y))

    return gdf


def run_temp_decrease(X, original_df, df, goal, name,simulation_mode):
    
    treatment_df = prepare_treatment_df(original_df, df)
    data = {"X": X.tolist()}
    response = requests.post(url, headers=headers, json=data)
    if response.ok:
        result = response.json()
        # print("Model response:", result["effect"])
    else:
        print("Request failed:", response.status_code, response.text)
    
    if simulation_mode == "Usar control deslizante":
        treatment_df['temp_decrease'] = -goal
        treatment_df['new_temp'] = treatment_df['LST'] + treatment_df['temp_decrease']
    else:
        treatment_df['new_temp'] = original_df['simulacion_temp']
        treatment_df['temp_decrease'] = treatment_df['new_temp'] - treatment_df['LST']
    
    treatment_effects = np.array(result["effect"])
    treatment_df[name] = treatment_df['temp_decrease'] / treatment_effects

    return treatment_df
def process_data_temp_to_trees(goal, simulation_mode, original_df, df, X, Y, T):

    name = 'pp_trees_increase'
    treatment_df = run_temp_decrease(X, original_df, df, goal, name,simulation_mode)

    treatment_df['pct_total_trees'] = treatment_df['Pct_CoberturaVeg'] + treatment_df[name]

    #add columns translated to square meters
    treatment_df['sqm_total_trees'] = (treatment_df['pct_total_trees']/100)*cell_area
    treatment_df['sqm_trees_increase'] = (treatment_df['pp_trees_increase']/100)*cell_area

    col_order = ['index', 'x', 'y', 'Pct_CoberturaVeg', 'Pct_Construccion', 'LST','temp_decrease', 
    'new_temp', 'pp_trees_increase','pct_total_trees', 'sqm_trees_increase','sqm_total_trees',]  # to match gdf trees

    treatment_df = treatment_df[col_order]

    gdf = gpd.GeoDataFrame(treatment_df, geometry=gpd.points_from_xy(treatment_df.x, treatment_df.y))

    return gdf