python-文件信息分析

具体要求

对本机的所有的文件信息数据进行搜索和收集，绘制数据分布的直方图。

1. 文件大小分布的直方图
2. 文件创建时间和修改时间的直方图
3. 文件类型的直方图

python 代码实现

import os
import time
import matplotlib.pyplot as plt
# import numpy as np
# from matplotlib import font_manager

# 初始化字符串为空
# str1 = ''
n_list = []
s_list = []
tc_list = []
tm_list = []
tp_list = []


def timeShift(timestamp):
    """
    时间戳转换时间
    :param timestamp: 时间戳
    :return: 固定格式的时间
    """
    time_local = time.localtime(timestamp)
    dt = time.strftime("%Y-%m-%d %H:%M:%S", time_local)
    return dt


def getFileInfo(fPath):
    """
    获取文件名称，大小，创建时间，修改时间以及文件类型
    返回值保存在列表中
    :param fPath: 想要获取的文件路径
    :return:
    """

    for root, dirs, files in os.walk(fPath):
        for file in files:
            if os.path.isfile(root + '//' + file):
                n_list.append(file)
                s_list.append(os.path.getsize(root + '//' + file))
                tc_list.append(timeShift(os.path.getctime(root + '//' + file)))
                tm_list.append(timeShift(os.path.getmtime(root + '//' + file)))
                tp_list.append(os.path.splitext(root + '//' + file)[1])
                # str1 = (file + '---' + str(os.path.getsize(root + '//' + file))
                #         + '---' + str(timeShift(os.path.getctime(root + '//' + file)))
                #         + '---' + str(timeShift(os.path.getmtime(root + '//' + file)))
                #         + '---' + str(os.path.splitext(root + '//' + file)[1]))
                # print(str1)
            else:
                return


# 调用函数对指定路径的文件进行遍历，获取文件信息


findPath = input("请输入需要查找的文件路径：")
getFileInfo(findPath)
# 对列表中保存的数据进行测试

# for i in n_list:
#     print(i)
#
# for i in s_list:
#     print(i)
#
# for i in tc_list:
#     print(i)
#
# for i in tm_list:
#     print(i)
#
# for i in tp_list:
#     print(i)

# 遍历数据进行判断
# sum1保存大小，sum2保存创建时间，sum3保存修改时间，sum4保存类型
sum1 = [0, 0, 0, 0]
sum2 = [0, 0, 0, 0]
sum3 = [0, 0, 0, 0]
sum4 = [0, 0, 0, 0, 0]


def size_Sort():
    """
    对文件大小进行分类
    """
    for i in s_list:
        if 0 <= i < 500:
            sum1[0] += 1
        elif 500 <= i < 2000:
            sum1[1] += 1
        elif 2000 <= i < 10000:
            sum1[2] += 1
        else:
            sum1[3] += 1


# 测试
# size_Sort()
# for i in sum1:
#     print(i)
def timec_Sort():
    """
    对文件创建时间时间进行分类
    """
    for i in tc_list:
        if '2019-06-01 00:00:00' <= i < '2019-07-01 00:00:00':
            sum2[0] += 1
        elif "2019-07-01 00:00:00" <= i < '2019-08-01 00:00:00':
            sum2[1] += 1
        elif "2019-08-01 00:00:00" <= i < '2019-09-01 00:00:00':
            sum2[2] += 1
        else:
            sum2[3] += 1


# 测试
# timec_Sort()
# for i in sum2:
#     print(i)


def timem_Sort():
    """
    对文件修改时间时间进行分类
    """
    for i in tm_list:
        if '2019-06-01 00:00:00' <= i < '2019-07-01 00:00:00':
            sum3[0] += 1
        elif "2019-07-01 00:00:00" <= i < '2019-08-01 00:00:00':
            sum3[1] += 1
        elif "2019-08-01 00:00:00" <= i < '2019-09-01 00:00:00':
            sum3[2] += 1
        else:
            sum3[3] += 1


# 测试
# timem_Sort()
# for i in sum2:
#     print(i)

def type_class():
    """
    对文件类型进行分类
    """

    for i in tp_list:
        if i == '.jpg' or i == '.png' or i == '.gif':
            sum4[0] += 1
        elif i == '.js' or i == '.css' or i == '.py' or i == '.htm':
            sum4[1] += 1
        elif i == '.txt' or i == '.xml' or i == '':
            sum4[2] += 1
        elif i == '.avi' or i == '.flv' or i == '.mp4' or i == '.wmv':
            sum4[3] += 1
        else:
            sum4[4] += 1


# 测试
# type_class()
# for i in sum4:
#     print(i)


def size_tu():
    """
    制作文件大小柱状图
    """
    size_Sort()
    name_list = [500, 2000, 10000, 100000]
    plt.bar(range(len(sum1)), sum1,
            tick_label=name_list)
    plt.show()


def timec_tu():
    """
    制作文件创建时间柱状图
    """
    timec_Sort()
    name_list = ['2019-06-01', '2019-07-01', '2019-08-01', '2019-09-01']
    plt.bar(range(len(sum2)), sum2,
            tick_label=name_list)
    plt.show()


def timem_tu():
    """
    制作文件修改时间柱状图
    """
    timem_Sort()
    name_list = ['2019-06-01', '2019-07-01', '2019-08-01', '2019-09-01']
    plt.bar(range(len(sum3)), sum3,
            tick_label=name_list)
    plt.show()


def type_tu():
    type_class()
    # 解决中文显示
    plt.rcParams['font.sans-serif'] = ['SimHei']
    plt.rcParams['axes.unicode_minus'] = False

    name_list = ['图片', '编程', '文本', '视频', '其他']
    plt.bar(range(len(sum4)), sum4,
            tick_label=name_list)
    plt.show()


size_tu()
timec_tu()
timem_tu()
type_tu()

运行结果

• 图1
• 

• 图2