speed up speech

main.py

@@ -74,17 +74,6 @@ async def qqreq(qq:QQ):
     print(qq.n2)
     return {"msg":"製作影片需要時間，請您耐心等候"} 
 
-@app.post("/make_video_req")
-async def make_video_req(req:request):
-    x = threading.Thread(target=make_video, args=(req.name, req.text_content, req.image_urls))
-    x.start()
-    return {"msg":"製作影片需要時間，請您耐心等候"} 
-
-@app.post("/make_anchor_video")
-async def make_anchor_video(req:request):
-    x = threading.Thread(target=anchor_video, args=(req.name, req.text_content, req.image_urls))
-    x.start()
-    return {"msg":"製作影片需要時間，請您耐心等候 稍後可以在www.choozmo.com:8168/"+req.name+".mp4 中觀看"} 
 
 @app.post("/make_anchor_video_v2")
 async def make_anchor_video_v2(req:request2):
@@ -104,90 +93,6 @@ def notify_group(msg):
     #print(r)
 
 
-def make_video(name, text_content, image_urls):
-    #save image
-    try:
-        os.mkdir(dir_photo+name)
-    except FileExistsError:
-        print("Directory " , dir_photo+name ,  " already exists")
-    img_num = 1
-    for imgu in image_urls:
-        im = Image.open(requests.get(imgu, stream=True).raw)
-        im.save(dir_photo+name+"/"+str(img_num)+".jpg")
-        img_num+=1
-    #save text
-    text_file = open(dir_text+name+".txt", "w")
-    text_file.write(text_content)
-    text_file.close()
-    print("text file made")
-    #make mp3
-    tts = zhtts.TTS() 
-    tts.text2wav(text_content,dir_sound+name+".mp3")
-    print("mp3 file made")
-    #make vido
-    t = openshot.Timeline(1280, 720, openshot.Fraction(30000, 1000), 44100, 2, openshot.LAYOUT_STEREO)
-    t.Open()
-    sound = openshot.FFmpegReader(dir_sound+name+".mp3")
-    sound.Open()
-    sound_track = openshot.Clip(sound)
-    t.AddClip(sound_track)
-    duration = sound.info.duration
-
-    w = openshot.FFmpegWriter("../html/"+name+".mp4")
-    w.SetAudioOptions(True, "libmp3lame", 44100, 2, openshot.LAYOUT_STEREO, 3000000)
-    w.SetVideoOptions(True, "libx264", openshot.Fraction(30000, 1000), 1280, 720,
-                        openshot.Fraction(1, 1), False, False, 3000000)
-    w.Open()
-    p_num = len(listdir(dir_photo+name))
-    clip_list = [None]*p_num
-    img_list = [None]*p_num
-    single_duration = int(duration/p_num)
-    frame_cnt = 0
-    for p in listdir(dir_photo+name):
-        img_list[frame_cnt] = openshot.QtImageReader(dir_photo+name+"/"+p)
-        clip_list[frame_cnt] = openshot.Clip(img_list[frame_cnt])
-        clip_list[frame_cnt].has_audio=openshot.Keyframe(0)
-        clip_list[frame_cnt].Layer(1)
-        clip_list[frame_cnt].Position(frame_cnt*single_duration) # start from time 0 sec
-        t.AddClip(clip_list[frame_cnt])
-        frame_cnt+=1
-
-    t_op = openshot.Timeline(1280, 720, openshot.Fraction(30000, 1000), 44100, 2, openshot.LAYOUT_STEREO)
-    t_op.Open()
-    op_vid = openshot.FFmpegReader(dir_video+"LOGO_OP.mp4")
-    op_vid.Open()
-    op_clip = openshot.Clip(op_vid)
-    t_op.AddClip(op_clip)
-    duration_op = op_vid.info.duration
-
-    for n in range(int(t_op.info.fps)*int(duration_op)):
-        f=t_op.GetFrame(n)
-        w.WriteFrame(f)
-
-    for n in range(int(t.info.fps)*int(duration)):
-        f=t.GetFrame(n)
-        w.WriteFrame(f)
-
-    t_ed = openshot.Timeline(1280, 720, openshot.Fraction(30000, 1000), 44100, 2, openshot.LAYOUT_STEREO)
-    t_ed.Open()
-    ed_vid = openshot.FFmpegReader(dir_video+"LOGO_ED.avi")
-    ed_vid.Open()
-    ed_clip = openshot.Clip(ed_vid)
-    t_ed.AddClip(ed_clip)
-    duration_ed = ed_vid.info.duration
-
-    for n in range(int(t_ed.info.fps)*int(duration_ed)):
-        f=t_ed.GetFrame(n)
-        w.WriteFrame(f)
-
-
-    t_ed.Close()
-    t_op.Close()
-    t.Close()
-    w.Close()
-    sound.Close()
-
-    notify_group(name+"的影片已經產生完成囉! www.choozmo.com:8168/"+name+".mp4")
 
 def cKey(r,g,b,fuzz):
     col=openshot.Color()
@@ -456,18 +361,6 @@ def split_by_pun(s):
     return res
 
 def generate_subtitle_image(name_hash,text_content):
-    '''
-    sub_len = 10
-    img_list = [None]*len(text_content)
-    for idx in range(len(text_content)):
-        img_list[idx] = []
-        for step in range(math.ceil(len(text_content[idx])/sub_len)):
-            sv_path = dir_subtitle + name_hash +'/'+str(idx)+ str(step) +'.png'
-            sub = text_content[idx][step*sub_len:step*sub_len+sub_len]
-            txt2image(sub, sv_path)
-            img_list[idx] += [{"count":len(sub),"path":sv_path}]
-            print(sub+':'+sv_path)
-    '''
     img_list = [None]*len(text_content)
     for idx in range(len(text_content)):
         img_list[idx]=[]
@@ -480,131 +373,6 @@ def generate_subtitle_image(name_hash,text_content):
     return img_list
 
 
-def anchor_video(name,text_content, image_urls):
-    m = hashlib.md5()
-    m.update(name.encode("utf-8"))
-    name_hash = m.hexdigest()
-
-    file_prepare(name, name_hash, text_content,image_urls)
-    call_achor_video(name_hash)
-
-    ck=cKey(0,254,0,150)
-    ck_anchor=cKey(0,255,1,320)
-    duration = 0
-    current_time = 0
-    current_time_photo = 0
-    #average layer level is 3
-    t = openshot.Timeline(1280, 720, openshot.Fraction(30000, 1000), 44100, 2, openshot.LAYOUT_STEREO)
-    t.Open()
-
-    bg_head = openshot.FFmpegReader(dir_video+"bg_head.avi")
-    bg_head.Open()
-    bg_head_clip = video_photo_clip(vid=bg_head,layer=2,position=current_time,end=bg_head.info.duration,ck=ck)
-    t.AddClip(bg_head_clip)
-
-    LOGO_OP = openshot.FFmpegReader(dir_video+"LOGO_OP.mp4")
-    LOGO_OP.Open()         # Open the reader
-    LOGO_OP_clip = video_photo_clip(vid=LOGO_OP,layer=4,position=current_time,end=bg_head.info.duration
-                    ,location_y=-0.03,scale_x=0.8,scale_y=0.71)
-    t.AddClip(LOGO_OP_clip)
-    duration+=bg_head.info.duration
-    current_time += bg_head.info.duration
-    current_time_photo += bg_head.info.duration
-
-    sound = openshot.FFmpegReader(dir_sound+name_hash+".mp3")
-    sound.Open()
-    sound_clip = openshot.Clip(sound)
-    sound_clip.Position(current_time)
-    t.AddClip(sound_clip)
-    duration += sound.info.duration
-
-    bg = openshot.FFmpegReader(dir_video+"bg.mp4")
-    bg.Open()
-    bg_times = math.floor(sound.info.duration/bg.info.duration)
-    left_time = sound.info.duration % bg.info.duration
-    bg_clip_list = [None] * bg_times
-    bg_list = [None] * bg_times
-    bg.Close()
-    for idx in range(bg_times):
-        bg_list[idx] = openshot.FFmpegReader(dir_video+"bg.mp4")
-        bg_list[idx].Open()
-        bg_clip_list[idx] = video_photo_clip(bg_list[idx],layer=2,position=current_time
-                ,end=bg_list[idx].info.duration,ck=ck)
-        t.AddClip(bg_clip_list[idx])
-        current_time += bg_list[idx].info.duration
-        bg_list[idx].Close()
-
-    bg_left = openshot.FFmpegReader(dir_video+"bg.mp4")
-    bg_left.Open()
-    bg_left_clip = video_photo_clip(bg_left,layer=2,position=current_time,end=left_time,ck=ck)
-    t.AddClip(bg_left_clip)
-    current_time += left_time
-    bg_left.Close()
-
-
-    p_num = len(listdir(dir_photo+name_hash))
-    photo_clip_list = [None]*p_num
-    img_list = [None]*p_num
-    photo_idx = 0
-    photo_duration = sound.info.duration/p_num
-    for p in listdir(dir_photo+name_hash):
-        img_list[photo_idx] = openshot.QtImageReader(dir_photo+name_hash+'/'+p)
-        photo_clip_list[photo_idx] = video_photo_clip(vid=img_list[photo_idx],layer=3
-                ,scale_x=0.81,scale_y=0.68,location_y=-0.03,position=current_time_photo,end=photo_duration
-                ,audio=False)
-        t.AddClip(photo_clip_list[photo_idx])
-        current_time_photo+=photo_duration
-        photo_idx+=1
-
-    current_time_anchor = bg_head.info.duration
-    anchor = openshot.FFmpegReader(dir_anchor+name_hash+".mp4")
-    anchor.Open()
-    anchor_times = math.floor(sound.info.duration/anchor.info.duration)
-    left_time = sound.info.duration % anchor.info.duration
-    anchor_clip_list = [None] * anchor_times
-    anchor_list = [None] * anchor_times
-    anchor.Close()
-    for idx in range(anchor_times):
-        anchor_list[idx] = openshot.FFmpegReader(dir_anchor+name_hash+".mp4")
-        anchor_list[idx].Open()
-        anchor_clip_list[idx] = video_photo_clip(vid=anchor_list[idx],layer=4,scale_x=0.65,scale_y=0.65,
-                location_x=0.35,location_y=0.22,position=current_time_anchor, end=anchor_list[idx].info.duration,
-                ck=ck_anchor,audio=False)
-        t.AddClip(anchor_clip_list[idx])
-        current_time_anchor += anchor_list[idx].info.duration
-        anchor_list[idx].Close()
-    anchor = openshot.FFmpegReader(dir_anchor+"peggy1_1.mp4")
-    anchor.Open()
-    anchor_left_clip = video_photo_clip(vid=anchor,layer=4,scale_x=0.65,scale_y=0.65,location_x=0.35,location_y=0.25
-            , position=current_time_anchor,end=left_time,ck=ck_anchor)
-    t.AddClip(anchor_left_clip)
-    anchor.Close()
-
-    title = openshot.QtImageReader(dir_title+name_hash+".png")
-    title.Open()         # Open the reader
-    title_clip = video_photo_clip(vid=title, layer=4, location_y=0.9,position=0,end=duration)
-    t.AddClip(title_clip)
-
-    ####start building
-    w = openshot.FFmpegWriter("../html/"+name_hash+".mp4")
-    w.SetAudioOptions(True, "libmp3lame", 44100, 2, openshot.LAYOUT_STEREO, 3000000)
-    w.SetVideoOptions(True, "libx264", openshot.Fraction(30000, 1000), 1280, 720,
-        openshot.Fraction(1, 1), False, False, 3000000)
-    w.Open()
-
-    #may change duration into t.info.duration
-    for n in range(int(t.info.fps)*int(duration)):
-        f=t.GetFrame(n)
-        w.WriteFrame(f)
-
-    t.Close()
-    w.Close()
-    sound.Close()
-
-    print("It's fine")
-    notify_group(name+"的影片已經產生完成囉! www.choozmo.com:8168/"+name_hash+".mp4")
-
-
 def anchor_video_v2(name,text_content, image_urls):
     m = hashlib.md5()
     m.update(name.encode("utf-8"))
@@ -689,7 +457,7 @@ def anchor_video_v2(name,text_content, image_urls):
         for sub_idx in range(len(sub_list[idx])):
             sub_img_list[idx][sub_idx] = openshot.QtImageReader(sub_list[idx][sub_idx]['path'])
             sub_img_list[idx][sub_idx].Open()
-            sub_duration = 0.25*sub_list[idx][sub_idx]['count']
+            sub_duration = 0.22*sub_list[idx][sub_idx]['count']
             sub_clip_list[idx][sub_idx] = video_photo_clip(vid=sub_img_list[idx][sub_idx], layer=6,location_x=0.069, location_y=0.89,position=main_timer+sub_timer,end=sub_duration)
             t.AddClip(sub_clip_list[idx][sub_idx])
             sub_img_list[idx][sub_idx].Close()