【Python】圧縮アルゴリズム（zlib、gzip、bz2）の速度・圧縮率の比較【サンプルあり】

import numpy as np
import pandas as pd
import os
import time
import pickle
import zlib
import gzip
import bz2
import gc

class StopWatch:
    '''
    sw = StopWatch()
    sw.start()
    print(f"*** 処理時間：{sw.stop()}")
    '''
    def start(self):
        self.__starttime = time.time()
        self.__time = 0
        return
    
    def stop(self):
        self.__time = time.time() - self.__starttime
        return self.__time
    
    def gettime(self):
        return self.__time

def objectsize(_obj):
    return round(_obj.__sizeof__()/(1024*1024),5)

def objectsizePrint(_obj):
    print(f" オブジェクトサイズ = {objectsize(_obj)}MB")

def compressAndDecompress(_object, _algo, _name):
    print(f"■{_name}")
    sw = StopWatch()

    try:
        sw.start()
        _compressed = _algo.compress(_object)
        print(f"{_name}.compress：{sw.stop()}")
        objectsizePrint(_compressed)
        _rec.append(sw.gettime())
        _rec.append(objectsize(_compressed))
    except:
        _rec.append("Err")
        _rec.append("Err")
        _compressed = ""

    try:
        sw.start()
        _decompressed = _algo.decompress(_compressed)
        print(f"{_name}.decompress：{sw.stop()}")
        objectsizePrint(_decompressed)
        _rec.append(sw.gettime())
    except:
        _rec.append("Err")

    del _compressed
    del _decompressed
    gc.collect() 


if __name__ == "__main__":
    sw = StopWatch()

    kekka = []
    _colN = 45

    for i in [3,30,300]:
        N = i*10000
        print(f"■{N/10000}万行{_colN}列----------")
        for j in range(5):
            _rec = []
            _rec.append(N)
            _rec.append(j+1)

            print(f"■■計測{j+1}回目----------")
            
            try:
                df = pd.DataFrame(np.ones((N,_colN)))
                df_pickled = pickle.dumps(df)
                objectsizePrint(df_pickled)
                _rec.append(objectsize(df_pickled))
            except:
                _rec.append("Err")

            compressAndDecompress(df_pickled, zlib, "zlib")
            compressAndDecompress(df_pickled, gzip, "gzip")
            compressAndDecompress(df_pickled, bz2, "bz2")

            kekka.append(_rec)

    #print(kekka)
    dfkekka = pd.DataFrame(kekka)
    dfkekka.columns = [
        'length', "times", 'fileSize', 
        "zlib.compress_time", "zlib.compress_size","zlib.decompress_time",
        "gzip.compress_time", "gzip.compress_size","gzip.decompress_time",
        "bz2.compress_time", "bz2.compress_size","bz2.decompress_time"
        ]
    dfkekka.to_csv('compose_decompose.csv')