Python 3 视频信息提取：自动生成 Excel 报告

import os
import re
import subprocess
import openpyxl
import math
from openpyxl.styles import Font, PatternFill, Alignment

# 视频格式
supported_formats = [
    '.mp4',
    '.avi',
    '.mkv',
    '.wmv',
    '.mov',
    '.flv',
    '.m2ts',
    '.ts',
    '.rm',
    '.rmvb',
    '.vob',
    '.3gp',
    '.webm',
    '.hdmov',
    '.mp4v',
    '.mpv4',
    '.divx',
    '.xvid',
    '.f4v',
    '.mpeg',
    '.asf',
    '.asx',
    '.m2t',
]

# 输出 Excel 路径
excel_folder = '.'

# 输出 Excel 文件名
excel_file = 'video_info.xlsx'

# 正则表达式匹配码率数据中的空格
pattern = re.compile(r'(\d+)\s+(\d+)')

# 创建 Excel 文件
wb = openpyxl.Workbook()
ws = wb.active
ws.title = 'Video Info'

# 锁定表头
ws.freeze_panes = 'A2'

# 设置表头样式
header_font = Font(bold=True, color='800080')  # 紫色
header_fill = PatternFill('solid', fgColor='C5E0B4')
header_alignment = Alignment(horizontal='center', vertical='center')
headers = [
    '文件名',
    '文件路径',
    '格式',
    '大小',
    '时长',
    '分辨率',
    '码率',
    '帧率',
    '音频码率',
    '音频采样率',
    '音频语言',
    '压缩比率',
    '文件总数',
]
for col, header in enumerate(headers, start=1):
    cell = ws.cell(row=1, column=col, value=header)
    cell.font = header_font
    cell.fill = header_fill
    cell.alignment = header_alignment

# 处理失败文件个数
failed_count = 0

# 总计处理文件数
total_count = 0

# 遍历所有视频文件
success_count = 0
row = 2 # 从第二行开始写入数据

def process_video(video_file):
    global success_count, row, failed_count
    try:
        # 获取文件名
        file_name = os.path.basename(video_file)

        # 获取文件大小
        size_str = convert_size(os.path.getsize(video_file))

        # 使用mediainfo获取视频信息 / 音频信息
        with subprocess.Popen(
            [
			'mediainfo',
                '--Inform=Video;%Format%|%Width%x%Height%|%BitRate/String%|%FrameRate%|%Duration/String3%',
                video_file
            ], stdout=subprocess.PIPE) as video_proc, 
            subprocess.Popen(
			[
			'mediainfo',
                '--Inform=Audio;%BitRate/String%|%SamplingRate/String%|%Language/String%',
                video_file
            ],
            stdout=subprocess.PIPE) as audio_proc:
            video_output = video_proc.communicate()[0].decode().strip()
            audio_output = audio_proc.communicate()[0].decode().strip()

        # 解析输出结果
        video_info = video_output.split('|')
        audio_info = audio_output.split('|')

        # 处理多音轨情况
        audio_bitrate, audio_sampling_rate, audio_lang = zip(*[(audio_info[i*3], 
        audio_info[i*3+1], audio_info[i*3+2]) for i in range(len(audio_info)//3)])

        # 时长取整
        duration = video_info[4].split('.')[0]
        h, m, s = duration.split(':')
        duration = f'{h}:{m}:{str(s).zfill(2)}'

        # 以分钟计算的时长
        duration_minutes = int(h) * 60 + int(m) + math.ceil(float(s)) / 60

        # 计算压缩比率
        ratio = round(duration_minutes / os.path.getsize(video_file) * 1000000000, 2)

        # 使用正则表达式替换码率数据中第1、2数字之间的空格
        bitrate = re.sub(pattern, r'\1\2', video_info[2])

        # 写入 Excel 文件
        write_cell(ws, row, 1, file_name)
        write_cell(ws, row, 2, os.path.dirname(video_file))  # 写入文件路径
        write_cell(ws, row, 3, video_info[0])
        write_cell(ws, row, 4, size_str)
        write_cell(ws, row, 5, duration)
        write_cell(ws, row, 6, video_info[1])
        write_cell(ws, row, 7, bitrate)
        write_cell(ws, row, 8, video_info[3])
        write_cell(ws, row, 9, ' / '.join(audio_bitrate))
        write_cell(ws, row, 10, ' / '.join(audio_sampling_rate))
        write_cell(ws, row, 11, ' / '.join(audio_lang))
        write_cell(ws, row, 12, ratio)
        write_cell(ws, 2, 13, total_count) # 写入文件总数

        # 打印进度
        success_count += 1
        row += 1

    except Exception as e:
        # 处理失败, 跳过并打印文件
        failed_count += 1
        print(f'处理文件 {video_file} 时失败：{e}')
        # 写入失败到err.txt
        with open('err.txt', 'a') as f:
            f.write(f'{video_file}\n')
        return False

    return True

def process_folder(video_folder):
    global total_count
    # 获取视频文件递归
    video_files = []
    for root, dirs, files in os.walk(video_folder):
        for file in files:
            if os.path.splitext(file)[1] in supported_formats:
                video_files.append(os.path.join(root, file))
    total_count = len(video_files)
    if total_count == 0:
        print('无视频文件或路径为空')
    else:
        for i, video_file in enumerate(video_files, start=1):
            if process_video(video_file):
                print(f'已处理 {success_count}/{total_count} 共计', end='\r')
            else:
                print(f'处理文件 {i}/{total_count} 时失败', end='\r')
        print(f'已处理 {success_count}/{total_count} 共计')

def convert_size(size):
    # 转换单位
    units = ['B', 'KiB', 'MiB', 'GiB']
    i = 0
    while size >= 1024 and i < len(units)-1:
        size /= 1024
        i += 1
    return f'{size:.2f} {units[i]}'

def write_cell(ws, row, col, value):
    # 写入Excel
    cell = ws.cell(row=row, column=col, value=value)
    if col == 13:
        cell.font = Font(bold=True, color='FF0000')
    if col in [10, 11, 13]:
        cell.alignment = Alignment(horizontal='left')
    else:
        cell.alignment = Alignment(horizontal='center')
    if col not in [3, 8, 9, 10, 11, 12, 13]:
        ws.column_dimensions[cell.column_letter].width = 11

# 文件路径
path = input('请输入文件路径：')
if os.path.isfile(path):
    # 1个文件 total_count += 1
    total_count = 1 
    process_video(path)
elif os.path.isdir(path):
    process_folder(path) 

if total_count > 0:
    # 保存 Excel 文件
    os.makedirs(excel_folder, exist_ok=True)
    excel_file_path = os.path.join(excel_folder, excel_file)
    wb.save(excel_file_path)

    # 打印处理失败文件个数和总计处理文件数
    print(
        f'成功处理 {success_count} 个文件, 失败 {failed_count} 个文件, 视频信息已保存到 {excel_file_path}'
    )