Select one row per model at the said percentile of price and print out price*1e6 and group size for each group

price-pct-by-model.py

#!/usr/bin/env python3

import argparse, math
import pandas as pd

def main():
    parser = argparse.ArgumentParser(
        description='Select one row per model at the said percentile of price and print out price*1e6 and group size for each group',
        formatter_class=argparse.ArgumentDefaultsHelpFormatter)
    parser.add_argument( '-p', '--percentile', type=float, default=25.0,
        help='Percentile (0-100)')
    parser.add_argument('csvfn', help='Input CSV filename')

    args = parser.parse_args()

    if not (0 <= args.percentile <= 100):
        parser.error('--percentile must be between 0 and 100')

    df = pd.read_csv(args.csvfn)
    df['Price'] = pd.to_numeric(df['Price'], errors='coerce')
    result = []
    for model, group in df.groupby('Model', sort=False):
        # print(model, len(group))
        g = group.sort_values('Price', kind='stable').reset_index()
        pos = math.floor((len(g) - 1) * args.percentile / 100)
        row = df.loc[g.loc[pos, "index"]].copy()
        row.iloc[-1] = len(group)
        result.append(row)
    result = pd.DataFrame(result)
    result.drop(columns=['Tier'], inplace=True)
    result.rename(columns={df.columns[-1]: 'Count'}, inplace=True)
    result['Price'] = (result['Price']*1e6).round().astype(int)
    print(result.to_csv(index=False), end='')

if __name__ == '__main__':
    main()