import pandas as pd
import numpy as np

def concatenate_data(weather_file, cons_file):
    df_wea = pd.read_csv(weather_file, sep=";")
    df_pow = pd.read_csv(cons_file, sep=";")

    df_wea.rename(index=str,
                    columns={
                        "Date UTC": "date",
                        "T° (C)": "temp",
                        "P (hPa)": "press",
                        "HR (%)": "humidity",
                        "P.rosée (°C)": "dew-point",
                        "Visi (km)": "visibl",
                        "Vt. moy. (km/h)": "wind-avg",
                        "Vt. raf. (km/h)": "wind-max",
                        "Vt. dir (°)": "wind-dir",
                        "RR 3h (mm)": "rain",
                        "Neige (cm)": "snow",
                        "Nebul. (octats)": "nebul"
                    },
                    inplace=True)

    qual = []
    quan = ["temp", "press", "humidity", "dew-point", "visibl", "wind-avg", "wind-max", "wind-dir", "rain", "snow", "nebul"]

    for i in qual:
        df_wea[i] = df_wea[i].fillna(value=df_wea[i].mode())
    for i in quan:
        df_wea[i] = df_wea[i].fillna(value=df_wea[i].median())

    df_wea["date"] = pd.to_datetime(df_wea["date"], format="%d/%m/%y %Hh%M")
    df_wea["weekday"] = df_wea["date"].dt.weekday
    df_wea["month"] = df_wea["date"].dt.month
    df_wea["year"] = df_wea["date"].dt.year
    df_wea["hour"] = df_wea["date"].dt.hour

    df_pow["date"] = pd.to_datetime(df_pow["date"], infer_datetime_format=True)

    df_pow = df_pow.drop_duplicates(subset="date").reset_index().set_index('date')
    ser_pow = df_pow["puissance"].resample("3H", label="right").sum()

    return df_wea.join(other=ser_pow, on="date")

concatenate_data("data/meteo_train.csv", "data/conso_train.csv")