python调用ffmpeg剪辑视频

#! /usr/bin/env python
# -*- coding: utf-8 -*-#
# -------------------------------------------------------------------------------
# Name:         剪辑视频
# Author:       yunhgu
# Date:         2021/10/21 14:08
# Description: 
# -------------------------------------------------------------------------------
import datetime
import os
import re
import subprocess
import time
from pathlib import Path
from traceback import format_exc

import openpyxl as px
from alive_progress import alive_bar
from ffmpy3 import FFmpeg
from moviepy.video.io.VideoFileClip import VideoFileClip

from log import log

logger = log("剪辑视频v2.3")
morning_start = 9 * 60 * 60
morning_end = 11 * 60 * 60

afternoon_start = 14 * 60 * 60
afternoon_end = 16 * 60 * 60


# 检查路径是否存在以及是否为空
def check_exist(path):
    return Path(path).exists() and path != ""


def load_excel(excel_path):
    try:
        open_excel = px.load_workbook(excel_path)
        sheet = open_excel.active
        return open_excel, sheet
    except Exception as e:
        logger.error(f"{excel_path}:加载excel失败,请确保后缀为.xlsx:{e}")


# 解析人工标注的视频开始时间为数字
def get_video_start_time(start_time):
    if re.match(r"d+:d+:d+", str(start_time)):
        x = time.strptime(str(start_time), '%H:%M:%S')
        second = datetime.timedelta(hours=x.tm_hour, minutes=x.tm_min, seconds=x.tm_sec).total_seconds()
        return second
    else:
        return -1


# 解析excel
def parse_excel(excel_path):
    category_date_name_time_dic = {}
    open_excel, sheet = load_excel(excel_path)
    for row_index in range(1, sheet.max_row + 1):
        video_name = sheet.cell(row_index, 1).value
        output_file_name = '_'.join(str(Path(video_name).name).split("_")[:3])
        date = sheet.cell(row_index, 2).value
        start_time = sheet.cell(row_index, 3).value
        if video_name is None and date is None and start_time is None:
            continue
        else:
            s_time = get_video_start_time(str(start_time).strip())
            if s_time == -1:
                logger.error(f"{excel_path}	行:{row_index} {start_time}格式不正确,跳过此条数据。")
            else:
                category_date_name_time_dic.setdefault(output_file_name, {})
                category_date_name_time_dic[output_file_name].setdefault(str(date), [])
                category_date_name_time_dic[output_file_name][str(date)].append(
                    {
                        "name": str(video_name).strip(),
                        "time": s_time
                    }
                )
    open_excel.close()
    return category_date_name_time_dic


# 匹配视频
def match_video_file(category_date_name_time_dic, input_path):
    for date_name_time_dic in category_date_name_time_dic.values():
        for name_time_list in date_name_time_dic.values():
            for name_time in name_time_list:
                name = name_time["name"]
                video_file_list = [file for file in input_path.rglob(Path(name).name)]
                if len(video_file_list) == 1:
                    name_time["video_file"] = video_file_list[0]
                else:
                    logger.error(f"{name}匹配不到对应的视频!")
                    name_time["video_file"] = None


def get_video_total_time(filename) -> float:
    video = VideoFileClip(str(filename))
    return video.duration


def get_cut_time(start_time_s, end_time_s, total_time):
    # print(f"{start_time_s}:{end_time_s}:{total_time}")
    if end_time_s < morning_start and start_time_s < morning_start:
        return 0, 0
    if start_time_s > morning_end and end_time_s < afternoon_start:
        return 0, 0
    if start_time_s > afternoon_end and end_time_s > afternoon_end:
        return 0, 0

    # 早上
    if start_time_s < morning_end:
        # print("morning")
        if start_time_s < morning_start:
            cut_start_time = morning_start - start_time_s
        else:
            cut_start_time = 0

        if end_time_s < morning_end:
            cut_end_time = total_time
        else:
            cut_end_time = total_time - (end_time_s - morning_end)
    # 下午
    else:
        # print("after")
        if start_time_s < afternoon_start:
            cut_start_time = afternoon_start - start_time_s
        else:
            cut_start_time = 0
        if end_time_s < afternoon_end:
            cut_end_time = total_time
        else:
            cut_end_time = total_time - (end_time_s - afternoon_end)
    return cut_start_time, cut_end_time


def generate_video_list(videos_name):
    with open("file_list.txt", encoding="utf-8", mode="w") as f:
        for name in videos_name:
            f.write(f"file {name}
")


def process_video(video_list, output_path, date, h_time="9_11"):
    videos_name = []
    try:
        output_file_name = ""
        for index, video in enumerate(video_list):
            file = video["video_file"]
            output_file_name = '_'.join(str(Path(file).name).split("_")[:3])
            start_time_s = video["time"]
            total_time = get_video_total_time(file)
            end_time_s = start_time_s + total_time
            if index == len(video_list) - 1:
                if h_time == "9_11" and end_time_s < morning_end or h_time == "14_16" and end_time_s < afternoon_end:
                    logger.error(f"{output_file_name}_{date}_{h_time}视频不足,生成视频时长不足2hr。")
            if index == 0:
                if h_time == "9_11" and start_time_s > morning_start or h_time == "14_16" and start_time_s > afternoon_start:
                    logger.error(f"{output_file_name}_{date}_{h_time}视频不足,生成视频时长不足2hr。")
            cut_start_time, cut_end_time = get_cut_time(start_time_s, end_time_s, total_time)
            if cut_start_time == 0 and cut_end_time == 0:
                continue
            # 裁剪视频
            FFmpeg(
                inputs={f"{file}": ['-ss', f"{cut_start_time}"]},
                outputs={f'output{index}.mp4': ['-to', f"{cut_end_time}", '-c', 'copy']},
                global_options=[
                    '-v', 'quiet',
                ]
            ).run(stdout=subprocess.PIPE)
            videos_name.append(f'output{index}.mp4')
        generate_video_list(videos_name)
        output_name = f"{output_file_name}_{date}_{h_time}.mp4"
        output_file_path = output_path.joinpath(f"{date}")
        output_file_path.mkdir(parents=True, exist_ok=True)
        output_video = str(output_file_path.joinpath(output_name))
        FFmpeg(
            inputs={"file_list.txt": ['-f', 'concat']},
            outputs={output_video: ['-acodec', 'copy', '-vcodec', 'copy', '-absf', 'aac_adtstoasc']},
            global_options=[
                '-v', 'quiet'
            ]
        ).run(stdout=subprocess.PIPE)
        for name in videos_name:
            os.remove(name)
    except Exception as e:
        logger.error(f"剪辑失败:{e}")
    finally:
        videos_name.append("file_list.txt")
        for name in videos_name:
            if os.path.exists(name):
                os.remove(name)


def convert(date, name_time_file_list, output_path):
    name_time_file_list = sorted(name_time_file_list, key=lambda a: a["time"])
    morning_video_list = []
    afternoon_video_list = []
    for name_time_file in name_time_file_list:
        if name_time_file["video_file"]:
            start_time_s = name_time_file["time"]
            if start_time_s < morning_end:
                morning_video_list.append(name_time_file)
            else:
                afternoon_video_list.append(name_time_file)
    if len(morning_video_list) > 0:
        process_video(morning_video_list, output_path, date, h_time="9_11")
    if len(afternoon_video_list) > 0:
        process_video(afternoon_video_list, output_path, date, h_time="14_16")


def main(input_path, excel_file_path, output_path):
    category_date_name_time_dic = parse_excel(excel_file_path)
    match_video_file(category_date_name_time_dic, input_path)
    # print(category_date_name_time_dic)

    for category, date_name_time_list_dic in category_date_name_time_dic.items():
        category_output_path = output_path.joinpath(category)
        category_output_path.mkdir(parents=True, exist_ok=True)
        with alive_bar(title=f"{category}:", total=len(date_name_time_list_dic)) as bar2:
            for date, name_time_file_list in date_name_time_list_dic.items():
                try:
                    convert(date, name_time_file_list, category_output_path)
                except Exception as e:
                    logger.error(f"{name_time_file_list}运行失败,跳过这个文件。{e}
{format_exc()}")
                finally:
                    bar2()


if __name__ == '__main__':
    while True:
        print("**** start    ****")
        input_folder = input("请输入包含视频文件夹:")
        excel_file = input("请输入excel文件:")
        output_folder = input("请输入结果保存文件夹:")

        # input_folder = r"C:UserspcDesktopvideo"
        # excel_file = r"F:任务2021王硕剪辑视频all_data.xlsx"
        # output_folder = r"F:任务2021王硕剪辑视频
esult"
        if check_exist(input_folder) and check_exist(output_folder):
            try:
                main(Path(input_folder), excel_file, Path(output_folder))
            except Exception as ee:
                logger.error(f"{format_exc()}:{ee}")
            print("**** finished ****")
            c = input("请输入q(不区分大小写)退出,按其他任意键继续!!!")
            if c.lower() == "q":
                break
        else:
            logger.error("输入的路径不存在,请检查后重新输入!!!")
            continue
不论你在什么时候开始,重要的是开始之后就不要停止。 不论你在什么时候结束,重要的是结束之后就不要悔恨。
原文地址:https://www.cnblogs.com/yunhgu/p/15465677.html