prepRain.py

"""
Code for precipitation data simulation preparation
Author: Vaclav Steinbach
Date: 15.08.2025
Dissertation Work
"""
import pandas as pd
import numpy as np
import os

"""
=-=-= CONFIG =-=-=
"""
campaign = "Amalie_2025-08-26_2025-09-07"
data_dir = "data/"+campaign+"/"
out_dir = "out/"
os.makedirs(out_dir, exist_ok=True)

# Select which file you want to process
# leave the other commented
filename = "LES2-S1" # Spruce forest probably..
csv_header = "LES2-S1" # this adresses the column var name in csv

filename = "LES1-S1" # Beech forest probably..
csv_header = "LES1-S1" # this adresses the column var name in csv
out_file = "rain_tree.in"

filename = "krajina_scintilometr"
out_file = "rain_free.in"
csv_header = "Scintilometr"

# Set the start and end date, set to none if you want all available data
start_date = "2025-09-05 00:00:00"  
end_date   = "2025-09-06 12:00:00"  

# start_date = None
# end_date   = None

# Set flag to true if you want to convert precipitation mm --> m
convert = True

"""
=-=-= TRANSFORM DATA =-=-=
"""
df = pd.read_csv(data_dir+filename+".csv")

# Rename columns
df = df.rename(columns={csv_header: "precip"})
df = df.rename(columns={csv_header: "precip"})

# Reformat time column to datetime
df["Time"] = pd.to_datetime(df["Time"])
# The raw data is sampled in 00:00:03 times
df["Time"] = df["Time"].dt.floor("10min")

# Select timerange
if start_date is not None and end_date is not None:
    df = df[(df["Time"] >= pd.to_datetime(start_date)) &
            (df["Time"] <= pd.to_datetime(end_date))]

# Convert Time to string in desired format
time_str = df["Time"].dt.strftime("%Y-%m-%dT%H:%M:%S")

# Save to file
np.savetxt(
    data_dir + "time.in",
    time_str,
    fmt="%s",  # important: save as string
    header="DTM",
    comments="# "
)
# Reformat precipitation into numeric
df["precip"] = pd.to_numeric(df["precip"], errors="coerce")
# Set datetime index and collapse duplicates (safe for resampling)
df = df.groupby("Time", as_index=True)["precip"].mean()

# Resample to 10-minute frequency and interpolate linearly
df = df.resample("10min").interpolate("linear")

# Reset index
df = df.reset_index()

# Convert precipitation from mm -> m
if convert:
    df["precip"] = df["precip"] / 1000.0 # m
else:
    df["precip"] = df["precip"] # mm

# Calculate seconds since start
start_time = df["Time"].iloc[0]
df["seconds"] = (df["Time"] - start_time).dt.total_seconds().astype(int)

# Reorder columns to seconds + precip
rain_arr = df[["seconds", "precip"]].to_numpy()

# Save to .in file
np.savetxt(
    data_dir+out_file,
    rain_arr,
    fmt="%d %.10f",
    header="seconds precipitation [m]",
    comments="# "
)

print("Data converted!")