AI_Anchor/OpenshotService/openshot_video_generator.py

from os import listdir
from os.path import isfile, isdir, join
import openshot
import threading
import zhtts
import os 
import urllib
from typing import List
import requests
from pydantic import BaseModel
from bs4 import BeautifulSoup
from PIL import Image,ImageDraw,ImageFont
import pyttsx3
import rpyc
import random
import re
import time
import math
import dataset
from datetime import datetime
from gtts import gTTS
import ffmpy

dir_sound = 'mp3_track/'
dir_photo = 'photo/'
dir_text = 'text_file/'
dir_video = 'video_material/'
dir_title = 'title/'
dir_subtitle = 'subtitle/'
dir_anchor = 'anchor_raw/'
tmp_video_dir = 'tmp_video/'
video_sub_folder = 'ai_anchor_video/'

dir_list = [dir_sound,dir_photo,dir_text,dir_video,dir_title,dir_subtitle,dir_anchor,tmp_video_dir]

def notify_group(msg):
    glist=['7vilzohcyQMPLfAMRloUawiTV4vtusZhxv8Czo7AJX8','WekCRfnAirSiSxALiD6gcm0B56EejsoK89zFbIaiZQD','1dbtJHbWVbrooXmQqc4r8OyRWDryjD4TMJ6DiDsdgsX','HOB1kVNgIb81tTB4Ort1BfhVp9GFo6NlToMQg88vEhh']
    for gid in glist:
        headers = {
                "Authorization": "Bearer " + gid,
                "Content-Type": "application/x-www-form-urlencoded"
        }
        params = {"message": msg}   
        r = requests.post("https://notify-api.line.me/api/notify",headers=headers, params=params)

def cKey(r,g,b,fuzz):
    col=openshot.Color()
    col.red=openshot.Keyframe(r)
    col.green=openshot.Keyframe(g)
    col.blue=openshot.Keyframe(b)
    return openshot.ChromaKey(col, openshot.Keyframe(fuzz))

def video_photo_clip(vid=None,layer=None, position=None, end=None
    ,scale_x=1,scale_y=1,location_x=0,location_y=0,ck=None,audio=True):
    clip = openshot.Clip(vid)
    clip.Layer(layer)
    clip.Position(position)
    clip.End(end)
    clip.scale_x=openshot.Keyframe(scale_x)
    clip.scale_y=openshot.Keyframe(scale_y)
    clip.location_x=openshot.Keyframe(location_x)
    clip.location_y=openshot.Keyframe(location_y)
    
    if ck!=None:
        clip.AddEffect(ck)
    if audio==True:
        clip.has_audio=openshot.Keyframe(1)
    else:
        clip.has_audio=openshot.Keyframe(0)
    return clip


def myunichchar(unicode_char):
        mb_string = unicode_char.encode('big5')
        try:
            unicode_char = unichr(ord(mb_string[0]) << 8 | ord(mb_string[1]))
        except NameError:
            unicode_char = chr(mb_string[0] << 8 | mb_string[1])
        return unicode_char

def get_url_type(url):
    req = urllib.request.Request(url, method='HEAD', headers={'User-Agent': 'Mozilla/5.0'})
    r = urllib.request.urlopen(req)
    contentType = r.getheader('Content-Type')
    return contentType
    
def make_dir(name_hash):
    for direct in dir_list:
        if not os.path.isdir(direct):
            os.mkdir(direct)
    try:
        os.mkdir(dir_photo+name_hash)
    except FileExistsError:
        print("~~~~~~Warning~~~~~~~~~Directory " , dir_photo+name_hash ,  " already exists")
    try:
        os.mkdir(dir_text+name_hash)
    except FileExistsError:
        print("~~~~~~Warning~~~~~~~~~Directory " , dir_text+name_hash ,  " already exists")
    try:
        os.mkdir(dir_sound+name_hash)
    except FileExistsError:
        print("~~~~~~Warning~~~~~~~~~Directory " , dir_sound+name_hash ,  " already exists")
    try:
        os.mkdir(dir_anchor+name_hash)
    except FileExistsError:
        print("~~~~~~Warning~~~~~~~~~Directory " , dir_anchor+name_hash ,  " already exists")
    try:
        os.mkdir(dir_subtitle+name_hash)
    except FileExistsError:
        print("~~~~~~Warning~~~~~~~~~Directory " , dir_subtitle+name_hash ,  " already exists")

def file_prepare(name, name_hash,text_content,image_urls,multiLang,lang='zh'):
    make_dir(name_hash)
    img_num = 1
    for imgu in image_urls:
        if get_url_type(imgu) =='video/mp4':
            r=requests.get(imgu)
            f=open(dir_photo+name_hash+"/"+str(img_num)+".mp4",'wb')
            for chunk in r.iter_content(chunk_size=255): 
                if chunk:
                    f.write(chunk)
            f.close()
        else:
            im = Image.open(requests.get(imgu, stream=True).raw)
            im= im.convert("RGB")
            im.save(dir_photo+name_hash+"/"+str(img_num)+".jpg")
        img_num+=1
    #save text
    txt_idx=0
    for txt in text_content:
        text_file = open(dir_text+name_hash+"/"+str(txt_idx)+".txt", "w")
        text_file.write(txt)
        text_file.close()
        txt_idx+=1
    print("text file made")
    #make mp3
    txt_idx = 0
    for txt in text_content:
        if lang!='zh' or multiLang==1:
            if lang!='zh':
                tts = gTTS(txt)
                tts.save(dir_sound+name_hash+"/"+str(txt_idx)+"raw.mp3")
            else:
                tts = gTTS(txt,lang='zh-tw')
                tts.save(dir_sound+name_hash+"/"+str(txt_idx)+"raw.mp3")
            #speed up 
            ff = ffmpy.FFmpeg(inputs={dir_sound+name_hash+"/"+str(txt_idx)+"raw.mp3": None}
                            , outputs={dir_sound+name_hash+"/"+str(txt_idx)+".mp3": ["-filter:a", "atempo=1.2"]})
            ff.run()
            os.remove(dir_sound+name_hash+"/"+str(txt_idx)+"raw.mp3")
        else:
            print('use zhtts')
            tts = zhtts.TTS() 
            tts.text2wav(txt,dir_sound+name_hash+"/"+str(txt_idx)+".mp3")
        txt_idx+=1
    print("mp3 file made")
    #make title as image
    txt2image_title(name, dir_title+name_hash+".png",lang)

def txt2image(content, save_target,lang='zh'):
    unicode_text = trim_punctuation(content)
    font = ''
    if lang=='zh':
        font = ImageFont.truetype(font="font/DFT_B7.ttc", size=38)
    else :
        font = ImageFont.truetype(font="font/arial.ttf", size=38)
    text_width, text_height = font.getsize(unicode_text)
    canvas = Image.new('RGBA', (700, 500), (255, 0, 0, 0) )
    draw = ImageDraw.Draw(canvas)
    text= unicode_text
    draw.text((5,5), text, (255, 255, 0), font)
    canvas.save(save_target, "PNG")

def txt2image_title(content, save_target, lang='zh'):
    unicode_text = trim_punctuation(content)
    font = ''
    if lang=='zh':
        font = ImageFont.truetype(font="font/DFT_B7.ttc", size=22)
    else :
        font = ImageFont.truetype(font="font/arial.ttf", size=22)
    text_width, text_height = font.getsize(unicode_text)
    canvas = Image.new('RGBA', (510, 500), (255, 0, 0, 0) )
    draw = ImageDraw.Draw(canvas)
    text= unicode_text
    draw.text((5,5), text, (17, 41, 167), font)
    canvas.save(save_target, "PNG")

def call_anchor(fileName,avatar):
    conn = rpyc.classic.connect("192.168.1.105",18812)
    ros = conn.modules.os 
    rsys = conn.modules.sys 
    fr=open(dir_sound+fileName+".mp3",'rb')# voice
    #warning!!!    file my be replaced by other process
    fw=conn.builtins.open('/tmp/output.mp3','wb')

    while True:
        b=fr.read(1024)
        if b:
            fw.write(b)
        else:
            break

    fr.close()
    fw.close()

    val=random.randint(1000000,9999999)
    ros.chdir('/home/jared/to_video')
    ros.system('./p'+str(avatar)+'.sh '+str(val)+' &')

    while True:
        print('waiting...')
        if ros.path.exists('/tmp/results/'+str(val)):
            break
        time.sleep(5)
        print('waiting...')

    fr=conn.builtins.open('/tmp/results/'+str(val)+'.mp4','rb')
    fw=open(dir_anchor+fileName+".mp4",'wb')
    while True:
        b=fr.read(1024)
        if b:
            fw.write(b)
        else:
            break

    fr.close()
    fw.close()

def trim_punctuation(s):
    pat_block = u'[^\u4e00-\u9fff0-9a-zA-Z]+';
    pattern = u'([0-9]+{0}[0-9]+)|{0}'.format(pat_block)
    res = re.sub(pattern, lambda x: x.group(1) if x.group(1) else u" " ,s)
    return res

def splitter(s):
    for sent in re.findall(u'[^!?，。\!\?]+[!?。\!\?]?', s, flags=re.U):
        yield sent

def split_by_pun(s):
    res = list(splitter(s))
    return res

def generate_subtitle_image(name_hash,text_content):
    img_list = [None]*len(text_content)
    for idx in range(len(text_content)):
        img_list[idx]=[]
        senList = split_by_pun(text_content[idx])
        for inner_idx in range(len(senList)):
            sv_path = dir_subtitle + name_hash +'/'+str(idx)+ str(inner_idx) +'.png'
            sub = senList[inner_idx]
            txt2image(sub,sv_path)
            img_list[idx]+=[{"count":len(sub),"path":sv_path}]
    return img_list

def generate_subtitle_image_ENG(name_hash,text_content):
    img_list = [None]*len(text_content)
    for idx in range(len(text_content)):
        sv_path = dir_subtitle + name_hash +'/'+str(idx)+'.png'
        sub = text_content[idx]
        txt2image(sub, sv_path,lang='eng')
        img_list[idx] = sv_path
    return img_list

def anchor_video_v2(name_hash,name,text_content, image_urls,multiLang,avatar):
    print(os.getcwd())
    print('sub image made')
    print(multiLang)
    file_prepare(name, name_hash, text_content,image_urls,multiLang)
    sub_list=generate_subtitle_image(name_hash,text_content)
    
    for fname in range(len(text_content)):
        call_anchor(name_hash+"/"+str(fname),avatar)
        print('step finish')
    print('called............................................')

    ck=cKey(0,254,0,270)
    ck_anchor=cKey(0,255,1,320)
    duration = 0
    #average layer level is 3
    t = openshot.Timeline(1280, 720, openshot.Fraction(30000, 1000), 44100, 2, openshot.LAYOUT_STEREO)
    t.Open()

    main_timer = 0
    
    LOGO_OP = openshot.FFmpegReader(dir_video+"LOGO_OP_4.mp4")
    LOGO_OP.Open()         # Open the reader
    LOGO_OP_clip = video_photo_clip(vid=LOGO_OP,layer=4,position=0,end=LOGO_OP.info.duration
                    ,location_y=-0.03,scale_x=0.8,scale_y=0.704)
    t.AddClip(LOGO_OP_clip)
    bg_head = openshot.FFmpegReader(dir_video+"complete_head_aispokesgirl.mp4")
    bg_head.Open()
    bg_head_clip = video_photo_clip(vid=bg_head,layer=2,position=0,end=LOGO_OP.info.duration,ck=ck)
    t.AddClip(bg_head_clip)
    main_timer += LOGO_OP.info.duration
    head_duration = LOGO_OP.info.duration
    bg_head.Close()
    LOGO_OP.Close()

    
    clip_duration=0
    photo_clip_list = [None]*len(text_content)
    img_list = [None]*len(text_content)
    anchor_clip_list = [None] * len(text_content)
    anchor_list = [None] * len(text_content)
    audio_clip_list = [None] * len(text_content)
    audio_list = [None] * len(text_content)
    sub_clip_list = [None] * len(text_content)
    sub_img_list = [None] * len(text_content)
    
    idx = 0
    for p in listdir(dir_photo+name_hash):
        
        anchor_list[idx] = openshot.FFmpegReader(dir_anchor+name_hash+"/"+str(idx)+".mp4")
        clip_duration = anchor_list[idx].info.duration
        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=0.25,position=main_timer, end=clip_duration,ck=ck_anchor,audio=False)
        t.AddClip(anchor_clip_list[idx])

        img_list[idx] = openshot.FFmpegReader(dir_photo+name_hash+'/'+p)
        img_list[idx].Open()
        photo_clip_list[idx] = video_photo_clip(vid=img_list[idx],layer=3
                ,scale_x=0.81,scale_y=0.68,location_y=-0.03,position=main_timer,end=clip_duration,audio=False)
        t.AddClip(photo_clip_list[idx])
        img_list[idx].Close()

        audio_list[idx] = openshot.FFmpegReader(dir_sound+name_hash+"/"+str(idx)+".mp3")
        audio_list[idx].Open()
        audio_clip_list[idx] = openshot.Clip(audio_list[idx])
        audio_clip_list[idx].Position(main_timer)
        audio_clip_list[idx].End(clip_duration)
        t.AddClip(audio_clip_list[idx])

        img_list[idx].Close()
        anchor_list[idx].Close()
        audio_list[idx].Close()
            
        sub_img_list[idx] = [None] * len(sub_list[idx])
        sub_clip_list[idx] = [None] * len(sub_list[idx])
        sub_timer = 0
        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 = 0.205*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()
            sub_timer += sub_duration
            print(sub_list[idx][sub_idx]['path'])
        main_timer += clip_duration
        idx+=1
    
    LOGO_ED = openshot.FFmpegReader(dir_video+"LOGO_ED.avi")
    LOGO_ED.Open()
    LOGO_ED_clip = video_photo_clip(vid=LOGO_ED,layer=4,position=main_timer,end=LOGO_ED.info.duration+2
                    ,location_x=0.005,location_y=-0.031
                    ,scale_x=0.8,scale_y=0.6825)
    t.AddClip(LOGO_ED_clip)
    ED_duration = LOGO_ED.info.duration
    LOGO_ED.Close()
    

    bg = openshot.FFmpegReader(dir_video+"complete_double_aispokesgirl.mp4")
    bg.Open()
    bg_times = math.floor(main_timer+ED_duration/bg.info.duration)
    left_time = (main_timer+ED_duration) % bg.info.duration
    bg_clip_list = [None] * bg_times
    bg_list = [None] * bg_times
    bg.Close()
    bg_timer = head_duration
    for idx in range(bg_times):
        bg_list[idx] = openshot.FFmpegReader(dir_video+"complete_double_aispokesgirl.mp4")
        bg_list[idx].Open()
        bg_clip_list[idx] = video_photo_clip(bg_list[idx],layer=2,position=bg_timer
                ,end=bg_list[idx].info.duration,ck=ck)
        t.AddClip(bg_clip_list[idx])
        bg_timer += bg_list[idx].info.duration
        bg_list[idx].Close()
    bg_left = openshot.FFmpegReader(dir_video+"complete_double_aispokesgirl.mp4")
    bg_left.Open()
    bg_left_clip = video_photo_clip(bg_left,layer=2,position=bg_timer,end=left_time,ck=ck)
    t.AddClip(bg_left_clip)
    bg_left.Close()

    title = openshot.QtImageReader(dir_title+name_hash+".png")
    title.Open()         # Open the reader
    title_clip = video_photo_clip(vid=title, layer=4,location_x=-0.047, location_y=0.801,position=0,end=head_duration+main_timer)
    t.AddClip(title_clip)

    ####start building
    w = openshot.FFmpegWriter(tmp_video_dir+name_hash+".mp4")
    w.SetAudioOptions(True, "aac", 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
    frames = int(t.info.fps)*int(head_duration+main_timer+ED_duration)
    for n in range(frames):
        f=t.GetFrame(n)
        w.WriteFrame(f)
        
    notify_group(name+"的影片已經產生完成囉! www.choozmo.com:8168/"+video_sub_folder+name_hash+".mp4")
    t.Close()
    w.Close()
    print("video at : www.choozmo.com:8168/"+video_sub_folder+name_hash+".mp4")


def anchor_video_eng(name_hash,name,text_content, image_urls,sub_titles,avatar):
    file_prepare(name, name_hash, text_content,image_urls,'eng')
    sub_list=generate_subtitle_image_ENG(name_hash,sub_titles)
    
    for fname in range(len(text_content)):
        call_anchor(name_hash+"/"+str(fname),avatar)
        print('step finish')
    print('called............................................')

    ck=cKey(0,254,0,270)
    ck_anchor=cKey(0,255,1,320)
    duration = 0
    #average layer level is 3
    t = openshot.Timeline(1280, 720, openshot.Fraction(30000, 1000), 44100, 2, openshot.LAYOUT_STEREO)
    t.Open()

    main_timer = 0
    #add logo
    LOGO_OP = openshot.FFmpegReader(dir_video+"LOGO_OP_4.mp4")
    LOGO_OP.Open()         # Open the reader
    LOGO_OP_clip = video_photo_clip(vid=LOGO_OP,layer=4,position=0,end=LOGO_OP.info.duration
                    ,location_y=-0.03,scale_x=0.8,scale_y=0.704)
    t.AddClip(LOGO_OP_clip)
    #add background video  (head is different)
    bg_head = openshot.FFmpegReader(dir_video+"complete_head_aispokesgirl.mp4")
    bg_head.Open()
    bg_head_clip = video_photo_clip(vid=bg_head,layer=2,position=0,end=LOGO_OP.info.duration,ck=ck)
    t.AddClip(bg_head_clip)
    
    main_timer += LOGO_OP.info.duration
    head_duration = LOGO_OP.info.duration
    bg_head.Close()
    LOGO_OP.Close()

    #prepare empty list 
    clip_duration=0
    photo_clip_list = [None]*len(text_content)
    img_list = [None]*len(text_content)
    anchor_clip_list = [None] * len(text_content)
    anchor_list = [None] * len(text_content)
    audio_clip_list = [None] * len(text_content)
    audio_list = [None] * len(text_content)
    sub_clip_list = [None] * len(text_content)
    #openshot image holder
    sub_img_list = [None] * len(text_content)
    
    idx = 0
    for p in listdir(dir_photo+name_hash):
        
        anchor_list[idx] = openshot.FFmpegReader(dir_anchor+name_hash+"/"+str(idx)+".mp4")
        clip_duration = anchor_list[idx].info.duration
        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=0.25,position=main_timer, end=clip_duration,ck=ck_anchor,audio=False)
        t.AddClip(anchor_clip_list[idx])
        #insert image 
        img_list[idx] = openshot.FFmpegReader(dir_photo+name_hash+'/'+p)
        img_list[idx].Open()
        photo_clip_list[idx] = video_photo_clip(vid=img_list[idx],layer=3
                ,scale_x=0.81,scale_y=0.68,location_y=-0.03,position=main_timer,end=clip_duration,audio=False)
        t.AddClip(photo_clip_list[idx])
        img_list[idx].Close()
        #insert audio (speech)
        audio_list[idx] = openshot.FFmpegReader(dir_sound+name_hash+"/"+str(idx)+".mp3")
        audio_list[idx].Open()
        audio_clip_list[idx] = openshot.Clip(audio_list[idx])
        audio_clip_list[idx].Position(main_timer)
        audio_clip_list[idx].End(clip_duration)
        t.AddClip(audio_clip_list[idx])
        #insert subtitle
        sub_img_list[idx] = openshot.QtImageReader(sub_list[idx])
        sub_img_list[idx].Open()
        sub_clip_list[idx] = video_photo_clip(vid=sub_img_list[idx], layer=6,location_x=0.069, location_y=0.89,position=main_timer,end=clip_duration)
        t.AddClip(sub_clip_list[idx])

        img_list[idx].Close()
        anchor_list[idx].Close()
        audio_list[idx].Close()
        sub_img_list[idx].Close()
            
        main_timer += clip_duration
        idx+=1
    
    LOGO_ED = openshot.FFmpegReader(dir_video+"ED_ENG.mp4")
    LOGO_ED.Open()
    LOGO_ED_clip = video_photo_clip(vid=LOGO_ED,layer=4,position=main_timer,end=LOGO_ED.info.duration+2
                    ,location_x=0.005,location_y=-0.031
                    ,scale_x=0.8,scale_y=0.6825)
    t.AddClip(LOGO_ED_clip)
    ED_duration = LOGO_ED.info.duration
    LOGO_ED.Close()
    

    bg = openshot.FFmpegReader(dir_video+"complete_double_aispokesgirl.mp4")
    bg.Open()
    bg_times = math.floor(main_timer+ED_duration/bg.info.duration)
    left_time = (main_timer+ED_duration) % bg.info.duration
    bg_clip_list = [None] * bg_times
    bg_list = [None] * bg_times
    bg.Close()
    bg_timer = head_duration
    for idx in range(bg_times):
        bg_list[idx] = openshot.FFmpegReader(dir_video+"complete_double_aispokesgirl.mp4")
        bg_list[idx].Open()
        bg_clip_list[idx] = video_photo_clip(bg_list[idx],layer=2,position=bg_timer
                ,end=bg_list[idx].info.duration,ck=ck)
        t.AddClip(bg_clip_list[idx])
        bg_timer += bg_list[idx].info.duration
        bg_list[idx].Close()
    bg_left = openshot.FFmpegReader(dir_video+"complete_double_aispokesgirl.mp4")
    bg_left.Open()
    bg_left_clip = video_photo_clip(bg_left,layer=2,position=bg_timer,end=left_time,ck=ck)
    t.AddClip(bg_left_clip)
    bg_left.Close()

    title = openshot.QtImageReader(dir_title+name_hash+".png")
    title.Open()         # Open the reader
    title_clip = video_photo_clip(vid=title, layer=4,location_x=-0.047, location_y=0.801,position=0,end=head_duration+main_timer)
    t.AddClip(title_clip)

    ####start building
    w = openshot.FFmpegWriter(tmp_video_dir+name_hash+".mp4")
    w.SetAudioOptions(True, "aac", 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
    frames = int(t.info.fps)*int(head_duration+main_timer+ED_duration)
    for n in range(frames):
        f=t.GetFrame(n)
        w.WriteFrame(f)
        
    notify_group(name+"(ENG)的影片已經產生完成囉! www.choozmo.com:8168/"+video_sub_folder+name_hash+".mp4")
    t.Close()
    w.Close()
    print("video at : www.choozmo.com:8168/"+video_sub_folder+name_hash+".mp4")
    #line notifs


class video_service(rpyc.Service):
    def exposed_call_video(self,name_hash,name,text_content, image_urls,multiLang,avatar):
        print('ML:'+str(multiLang))
        anchor_video_v2(name_hash,name,text_content, image_urls,multiLang,avatar)
    def exposed_call_video_eng(self,name_hash,name,text_content, image_urls,sub_titles,avatar):
        anchor_video_eng(name_hash,name,text_content, image_urls,sub_titles,avatar)
    def exposed_call_videoGen(self,name_hash,name,text_content, image_urls,multiLang,avatar):
        print('ML:'+str(multiLang))
        anchor_video_v2(name_hash,name,text_content, image_urls,multiLang,avatar)



from rpyc.utils.server import ThreadedServer
t = ThreadedServer(video_service, port=8858)
print('service started')
t.start()