利用python通过字幕文件.srt来实现对视频片段的截取（2）_根据字幕中的关键词剪切视频

 
"""
        用于获取多个关键词的视频摘要
        本代码还是有缺陷的，因为选择了几个for循环，导致时间还是挺慢的。
"""
 
import cv2 as cv
import numpy as np
import re
import os
from PIL import Image
video_file = cv.VideoCapture(r"E:\900秒嫦娥落月.mp4")
fps = video_file.get(5)
print("该视频的帧为{}".format(fps))
#==================================================================================#
"""
    实验阶段：该内容是用来获取想要寻找的关键字段，在视频中时间位置，并进行截取视频
    最终阶段：进行文本匹配，谁的相似度最高，就选择哪一段进行截取，
            而在此任务中涉及到的知识很多：文本相似度匹配、近义词训练、如何将txt文件的字幕和时间分割再还原等等问题
"""
#name=input("请输入提取内容：")
# name=input("请输入你所需要寻找的关键字段或者关键内容:")
# with open("CHS_嫦娥奔月.txt","r",encoding="utf-8") as f:
#     flie_data=f.readlines()
#     for i,readline in enumerate(flie_data):
#         #print(readline)
#         # matchObj=re.match(name,readline)#匹配时，如果第一个字符不匹配，就会发生匹配失败
#         matchObj = re.search(name, readline)
#         print(matchObj)
#         if matchObj is not None:
#             print("匹配字符所在的位置{}".format(i))
#             print("所在的时间段为{}".format(flie_data[i-1]))
#             break
list_time=[]
list_frames_time=[]
name_all=input("请输入需要提取的关键词（用|分开）:")
name_all=name_all.split("|")
# name_all=["燃料","失败"]
with open(r"D:\PycharmProjects\pythonProject\CHS_嫦娥奔月.txt","r",encoding="utf-8") as f:
    flie_data=f.readlines()
    for name in name_all:
        for i,readline in enumerate(flie_data):
            #print(readline)
            # matchObj=re.match(name,readline)#匹配时，如果第一个字符不匹配，就会发生匹配失败
            matchObj = re.search(name, readline)
            print(matchObj)
            if matchObj is not None:
                list_time.append(flie_data[i-1])
                print("匹配字符所在的位置{}".format(i))
                print("所在的时间段为{}".format(flie_data[i-1]))
list_time.sort()#将时间进行排序（从小到大）
print("需要截取的视频时间",list_time)
# =============================================================================#
# 首先弄人工输入，并将帧和时间进行转换
# start_frame = input("请输入开始帧数,(输入格式为时:分:秒,毫秒或者帧):")
for start_frame in list_time:
    #start_frame=flie_data[i-1]
    if "-->" in start_frame:
        start_frame_1 = start_frame.split(" --> ")
        start_frame = start_frame_1[0]
        end_frame = start_frame_1[1]
        print("起始时间-->终止时间")
    else:
        end_frame = input("请结束开始帧数,(输入格式为时:分:秒,毫秒或者帧):")
        """
        如果得到了是时间格式就进行转换成帧格式
        输入的是时:分:秒.毫秒
        1时=60分=3600秒=3600000毫秒
        如果得到的是帧格式，那正好不用转换
        """
        # start_frame_spilt=start_frame.split(":")
    # ==============================================================================#
    if ":" in start_frame:
        print("输入的是时间制！")
        name_str = start_frame.split(":")
        start_frame = (int(name_str[0]) * 3600 + int(name_str[1]) * 60 + float(name_str[2].replace(",", "."))) * fps
        list_frames_time.append(start_frame)
    else:
        print("输入的值十进制（帧）！")
        start_frame = int(start_frame)
    if ":" in end_frame:
        print("输入的是时间制！")
        name_str = end_frame.split(":")
        end_frame = (int(name_str[0]) * 3600 + int(name_str[1]) * 60 + float(name_str[2].replace(",", "."))) * fps
        list_frames_time.append(end_frame)
    else:
        print("输入的值十进制（帧）！")
        end_frame = int(end_frame)
        # ==============================================================================================#
        """
            如果前后的时间比较小，就是帧数小于1，那么无法截到图，所以直接截前后5帧来进行填充
        """
    if (end_frame - start_frame) < 1:
        start_frame = start_frame - 5
        end_frame = end_frame + 5
    else:
        pass
print(list_frames_time)
count_frame=0
for i in range(0,len(list_frames_time),2):
    start_frame=list_frames_time[i]
    end_frame=list_frames_time[i+1]
    success, frames = video_file.read()
    while success:
        success, frames = video_file.read()
        count_frame += 1
        if count_frame >= start_frame and count_frame <= end_frame:
            print("截取第{}帧".format(count_frame))
            cv.imwrite(r"D:\PycharmProjects\pythonProject\feiji\frames\image{}.jpg".format(count_frame), frames)
        if count_frame > end_frame:
            break
 
print("开始合成视频")
im_list = os.listdir("D:/PycharmProjects/pythonProject/feiji/frames/")
#im_list.sort(key=lambda x: int(x.replace("frame", "").split('.')[0]))  # 最好再看看图片顺序对不
img = Image.open(os.path.join("D:/PycharmProjects/pythonProject/feiji/frames/", im_list[0]))
img_size = img.size  # 获得图片分辨率，im_dir文件夹下的图片分辨率需要一致
 
# fourcc = cv2.cv.CV_FOURCC('M','J','P','G') #opencv版本是2
fourcc = cv.VideoWriter_fourcc(*'MJPG')  # opencv版本是3
videoWriter = cv.VideoWriter(r"D:\PycharmProjects\pythonProject\feiji\frames\test.avi", fourcc, fps, img_size)
# count = 1
for i in im_list:
    im_name = os.path.join("D:/PycharmProjects/pythonProject/feiji/frames/" + i)
    frame = cv.imdecode(np.fromfile(im_name, dtype=np.uint8), -1)
    videoWriter.write(frame)
    # count+=1
    # if (count == 200):
    #     print(im_name)
    #     break
videoWriter.release()
print('finish')