dashboard-temp-final / data_utils.py
General-ProjectsByIF's picture
Update data_utils.py
990b611 verified
Raw
History Blame Contribute Delete
4.81 kB
import pandas as pd
import numpy as np
import geopandas as gpd
import requests
url = "https://Projects-by-IF-model-temp-api.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