Sort cases

[zxdawn]

For the generated fresh_lightning_cases.csv or no_lightning_cases.csv, we sometimes drop multiple rows because it's bad cases. Then, we need to sort case and reindex

In case anyone is interested, I post here.

import os
import pandas as pd
from itertools import groupby
from s5p_lno2_utils import Config


def main():
    df = pd.read_csv(os.path.join(data_dir, filename))
    df['case'] = df.groupby(['case']).ngroup()
    print(df)
    df.to_csv(os.path.join(data_dir, 'no_lightning_cases.csv'), index=False)


    # get the basenames of the filelist
    basenames = list(sorted(map(os.path.basename, df['filename'])))

    # get the dates of consecutive swaths for downloading CTMANA files
    dates = []
    for _, g in groupby(enumerate(basenames), lambda k: int(k[1].split('_')[-3]) - k[0]):
        dates.extend([v.split('_')[-5][:8] for _, v in g])

    # export dates for downloading CTMANA files
    dates = list(sorted(set(dates)))
    df_dates = pd.DataFrame(dates, columns=['dates'])
    print(df_dates)
    print(f'Saved the Dates(yyyymmdd) DataFrame to {savename}')
    df_dates.to_csv('ctmana_dates.csv', index=False, header=False)


if __name__ == '__main__':
    # read config file
    cfg = Config('settings.txt')

    # directory where cases file are saved
    data_dir = cfg['output_data_dir']
    filename = 'no_lightning_cases_check.csv'

    main()