def make_video_main(video_n=1):
img_path='/media/xiaolong/datapool/data/downloads/nvshens/'
result_video_path='/media/xiaolong/datapool/data/downloads/test/'
audio_files_path='/media/xiaolong/datapool/data/youtube_video/audio/'
audio_files=glob.glob(audio_files_path+'*.mp3')
audio_n=randint(0, len(audio_files))
audio_file_path=audio_files[audio_n]#'/media/xiaolong/datapool/data/youtube_video/audio/'gudian_'+str(audio_n)+'.mp3''
video_file_path=result_video_path+'result_stocking.mp4'.replace('.mp4',str(video_n)+'.mp4')
mp4_name_str=['stocking_a.mp4','stocking_b.mp4','stocking_c.mp4']
for i in range(len(mp4_name_str)):
mp4_name_str[i]=mp4_name_str[i].replace('.mp4',str(video_n)+'.mp4')
for mp4_name in mp4_name_str:
if not os.path.exists(result_video_path+mp4_name):
generate_video_shell(img_path,result_video_path+mp4_name)
file_names=''
for mp4_name in mp4_name_str:
if file_names=='':
file_names=' -i '+result_video_path+mp4_name
else:
file_names+=' -i '+result_video_path+mp4_name
if not os.path.exists(video_file_path):
cmd_str='ffmpeg'+file_names
os.system(cmd_str+' -filter_complex "[0:v][1:v]concat=n=2:v=1:a=0[slideshow]" -map "[slideshow]" ' +video_file_path)
video_audio_dir=video_file_path.replace('.mp4','_audio.mp4')
if not os.path.exists(video_audio_dir):
ops.combine_audio_video_cmd(audio_file_path, video_file_path, video_audio_dir)
def generate_news_video_function(pickle_file_path,lan='chinese'):
#pickle_file_path='/media/xiaolong/datapool/data/youtube/news/mitbbs/news_database/31918997.p'
with open(pickle_file_path, 'rb') as fp:
data = pickle.load(fp)
#audio_video_path=data['images_set']
folder_base_path=data['images_set']
#import IPython
#IPython.embed()
news_audio_video_path=folder_base_path.split('image')[0]+'video/'
news_img_path=folder_base_path#+'images/'
if not os.path.exists(news_audio_video_path):
os.makedirs(news_audio_video_path)
release_video_path=folder_base_path.split('image')[0]+'release_video/'
if not os.path.exists(release_video_path):
os.makedirs(release_video_path)
if not os.path.exists(news_img_path):
os.makedirs(news_img_path)
article_title=data['title']
result_mp4_file=news_audio_video_path+article_title.replace(' ','_')+'.mp4'
v_a_file=result_mp4_file.replace('.mp4','_audio.mp4')
v_a_txt_file=result_mp4_file.replace('.mp4','_audio_txt.mp4')
audio_result_file='/media/xiaolong/datapool/data/youtube/news/sina_news/bana_news.mp3'
text_on_video=data['text_article']
#generate_mp3_function(news_str,audio_result_file)
subtitle_file_path=result_mp4_file.replace('.mp4','.txt')
batch_num=generate_text_fun(data,result_txt_file=subtitle_file_path)
#img_files,one_min_photos,num_video,num_photos_needed=photo_calculation(news_img_path,audio_result_file,article_title)
img_files=photo_calculation(news_img_path,audio_result_file,article_title)
if not os.path.exists(result_mp4_file):
generate_sub_video_shell(img_files,result_mp4_file)
if not os.path.exists(v_a_file):
ops.combine_audio_video_cmd(audio_result_file, result_mp4_file, v_a_file)
#import IPython
#IPython.embed()
if not os.path.exists(v_a_txt_file):
generate_video_text_fun(input_video=v_a_file,output_video=v_a_txt_file,text_file=subtitle_file_path, batch_num=batch_num)
#return
mp4_new_name=result_mp4_file.replace('video',data['title'].decode('utf-8'))
#shutil.copyfile(result_mp4_file, mp4_new_name.encode('utf-8') )
#if not os.path.exists(mp4_new_name):
# os.makedirs(mp4_new_name)
add_watermark_to_video(v_a_txt_file)
def video_add_audio_sub_title_fun(audio_result_file,result_mp4_file,subtitle_path):
sub_title_video=result_mp4_file.replace('.mp4','_title.mp4')
video_audio_dir=result_mp4_file.replace('.mp4','_audio.mp4')
if not os.path.exists(sub_title_video):
os.system('ffmpeg -i '+result_mp4_file+' -vf ass='+subtitle_path+' '+sub_title_video)
if not os.path.exists(video_audio_dir):
ops.combine_audio_video_cmd(audio_result_file, result_mp4_file, video_audio_dir)
def generate_video_main(audio_file_path,video_file_path):
video_audio_dir=video_file_path.replace('.mp4','_v.mp4')
ops.combine_audio_video_cmd(audio_file_path, video_file_path, video_audio_dir)
def generate_news_video_function(lan='chinese'):
pickle_file_path='/media/xiaolong/datapool/data/youtube/news/mitbbs/31919397.p'
with open(pickle_file_path, 'rb') as fp:
data = pickle.load(fp)
audio_video_path=data['images_set']
result_mp4_file=audio_video_path+'video.mp4'
audio_result_file=audio_video_path+'audio.mp3'
subtitle_path=audio_video_path+'/subtitle/audio.ass'
#generate mp3 files
news_str=data['text_article']
generate_mp3_function(news_str,audio_result_file)
#import IPython
#IPython.embed()
if not os.path.exists(subtitle_path):
wave_file=audio_result_file.replace('.mp3','.wav')
os.system('ffmpeg -i '+audio_result_file+' '+wave_file)
autosubtitle_chinese.generate_subtitle(wave_file,news_str,language='zh')
return
#generate_news_video_shell(audio_video_path,result_mp4_file,style='animation')
article_title=data['title']
audio = MP3(audio_result_file)
audio_length_second=audio.info.length
num_photos_needed=int(audio_length_second/10)+1
#generate mp4 video files
img_files1=glob.glob(audio_video_path+'*.jpeg')
img_files2=glob.glob(audio_video_path+'*.jpg')
img_files3=glob.glob(audio_video_path+'*.png')
img_files=img_files1+img_files2+img_files3
img_files = sorted(img_files,key=lambda x: int(os.path.splitext(x.split('/')[-1])[0]))
#shuffle(img_files)
time_per_photo=10
total_n_im_now=len(img_files)
if total_n_im_now<num_photos_needed:
#download_google_image(key_word=article_title.encode('ascii', 'ignore').decode('ascii'),output_path=audio_video_path)
#audio_video_main.download_google_image(key_word=article_title,output_path=audio_video_path)
os.system('python3 '+baidu_download+'baidu_image_download.py --key_word '+article_title.encode('utf-8')+' --image_save_path ' +audio_video_path)
audio_video_main.image_filter_function(audio_video_path)
one_min_photos=60/time_per_photo
num_video=num_photos_needed/one_min_photos
sub_video_name_str=[]
for i in range(0,num_video):
sub_result_video_path=result_mp4_file.replace('.mp4',str(i)+'.mp4')
if not i==num_video-1:
images_batch=img_files[i*one_min_photos:(i+1)*one_min_photos]
else:
images_batch=img_files[i*one_min_photos:num_photos_needed]
if not os.path.exists(sub_result_video_path):
generate_sub_video_shell(images_batch,sub_result_video_path)
sub_video_name_str.append(sub_result_video_path)
file_names=''
for mp4_name in sub_video_name_str:
if file_names=='':
file_names=' -i '+mp4_name
else:
file_names+=' -i '+mp4_name
if not os.path.exists(result_mp4_file):
cmd_str='ffmpeg'+file_names
os.system(cmd_str+' -filter_complex "[0:v][1:v]concat=n=2:v=1:a=0[slideshow]" -map "[slideshow]" ' +result_mp4_file)
news_subtitle=news_str.split(u'\uff0c')
# sub_title=[]
# for sentence_sub in news_subtitle:
# dot_sen=sentence_sub.split(u'\u3002')
# for dot_sen_sub in dot_sen:
# sub_title.append(dot_sen_sub)
print(re.split('\u3002, \uff0c',news_str))
sub_title_video=result_mp4_file.replace('.mp4','_title.mp4')
os.system('ffmpeg -i '+result_mp4_file+' -vf ass='+subtitle_path+' '+sub_title_video)
video_audio_dir=result_mp4_file.replace('.mp4','_audio.mp4')
if not os.path.exists(video_audio_dir):
ops.combine_audio_video_cmd(audio_result_file, sub_title_video, video_audio_dir)
