File size: 8,269 Bytes
5f685fd |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 |
import datetime
import os
import re
import shutil
from abc import abstractmethod
from shortGPT.audio import audio_utils
from shortGPT.audio.audio_duration import get_asset_duration
from shortGPT.audio.voice_module import VoiceModule
from shortGPT.config.asset_db import AssetDatabase
from shortGPT.config.languages import Language
from shortGPT.editing_framework.editing_engine import (EditingEngine,
EditingStep)
from shortGPT.editing_utils import captions, editing_images
from shortGPT.editing_utils.handle_videos import extract_random_clip_from_video
from shortGPT.engine.abstract_content_engine import AbstractContentEngine
from shortGPT.gpt import gpt_editing, gpt_translate, gpt_yt
class ContentShortEngine(AbstractContentEngine):
def __init__(self, short_type: str, background_video_name: str, background_music_name: str, voiceModule: VoiceModule, short_id="",
num_images=None, watermark=None, language: Language = Language.ENGLISH,):
super().__init__(short_id, short_type, language, voiceModule)
if not short_id:
if (num_images):
self._db_num_images = num_images
if (watermark):
self._db_watermark = watermark
self._db_background_video_name = background_video_name
self._db_background_music_name = background_music_name
self.stepDict = {
1: self._generateScript,
2: self._generateTempAudio,
3: self._speedUpAudio,
4: self._timeCaptions,
5: self._generateImageSearchTerms,
6: self._generateImageUrls,
7: self._chooseBackgroundMusic,
8: self._chooseBackgroundVideo,
9: self._prepareBackgroundAssets,
10: self._prepareCustomAssets,
11: self._editAndRenderShort,
12: self._addYoutubeMetadata
}
@abstractmethod
def _generateScript(self):
self._db_script = ""
def _generateTempAudio(self):
if not self._db_script:
raise NotImplementedError("generateScript method must set self._db_script.")
if (self._db_temp_audio_path):
return
self.verifyParameters(text=self._db_script)
script = self._db_script
if (self._db_language != Language.ENGLISH.value):
self._db_translated_script = gpt_translate.translateContent(script, self._db_language)
script = self._db_translated_script
self._db_temp_audio_path = self.voiceModule.generate_voice(
script, self.dynamicAssetDir + "temp_audio_path.wav")
def _speedUpAudio(self):
if (self._db_audio_path):
return
self.verifyParameters(tempAudioPath=self._db_temp_audio_path)
self._db_audio_path = audio_utils.speedUpAudio(
self._db_temp_audio_path, self.dynamicAssetDir+"audio_voice.wav")
def _timeCaptions(self):
self.verifyParameters(audioPath=self._db_audio_path)
whisper_analysis = audio_utils.audioToText(self._db_audio_path)
self._db_timed_captions = captions.getCaptionsWithTime(
whisper_analysis)
def _generateImageSearchTerms(self):
self.verifyParameters(captionsTimed=self._db_timed_captions)
if self._db_num_images:
self._db_timed_image_searches = gpt_editing.getImageQueryPairs(
self._db_timed_captions, n=self._db_num_images)
def _generateImageUrls(self):
if self._db_timed_image_searches:
self._db_timed_image_urls = editing_images.getImageUrlsTimed(
self._db_timed_image_searches)
def _chooseBackgroundMusic(self):
self._db_background_music_url = AssetDatabase.get_asset_link(self._db_background_music_name)
def _chooseBackgroundVideo(self):
self._db_background_video_url = AssetDatabase.get_asset_link(
self._db_background_video_name)
self._db_background_video_duration = AssetDatabase.get_asset_duration(
self._db_background_video_name)
def _prepareBackgroundAssets(self):
self.verifyParameters(
voiceover_audio_url=self._db_audio_path,
video_duration=self._db_background_video_duration,
background_video_url=self._db_background_video_url, music_url=self._db_background_music_url)
if not self._db_voiceover_duration:
self.logger("Rendering short: (1/4) preparing voice asset...")
self._db_audio_path, self._db_voiceover_duration = get_asset_duration(
self._db_audio_path, isVideo=False)
if not self._db_background_trimmed:
self.logger("Rendering short: (2/4) preparing background video asset...")
self._db_background_trimmed = extract_random_clip_from_video(
self._db_background_video_url, self._db_background_video_duration, self._db_voiceover_duration, self.dynamicAssetDir + "clipped_background.mp4")
def _prepareCustomAssets(self):
self.logger("Rendering short: (3/4) preparing custom assets...")
pass
def _editAndRenderShort(self):
self.verifyParameters(
voiceover_audio_url=self._db_audio_path,
video_duration=self._db_background_video_duration,
music_url=self._db_background_music_url)
outputPath = self.dynamicAssetDir+"rendered_video.mp4"
if not (os.path.exists(outputPath)):
self.logger("Rendering short: Starting automated editing...")
videoEditor = EditingEngine()
videoEditor.addEditingStep(EditingStep.ADD_VOICEOVER_AUDIO, {
'url': self._db_audio_path})
videoEditor.addEditingStep(EditingStep.ADD_BACKGROUND_MUSIC, {'url': self._db_background_music_url,
'loop_background_music': self._db_voiceover_duration,
"volume_percentage": 0.11})
videoEditor.addEditingStep(EditingStep.CROP_1920x1080, {
'url': self._db_background_trimmed})
videoEditor.addEditingStep(EditingStep.ADD_SUBSCRIBE_ANIMATION, {'url': AssetDatabase.get_asset_link('subscribe animation')})
if self._db_watermark:
videoEditor.addEditingStep(EditingStep.ADD_WATERMARK, {
'text': self._db_watermark})
caption_type = EditingStep.ADD_CAPTION_SHORT_ARABIC if self._db_language == Language.ARABIC.value else EditingStep.ADD_CAPTION_SHORT
for timing, text in self._db_timed_captions:
videoEditor.addEditingStep(caption_type, {'text': text.upper(),
'set_time_start': timing[0],
'set_time_end': timing[1]})
if self._db_num_images:
for timing, image_url in self._db_timed_image_urls:
videoEditor.addEditingStep(EditingStep.SHOW_IMAGE, {'url': image_url,
'set_time_start': timing[0],
'set_time_end': timing[1]})
videoEditor.renderVideo(outputPath, logger= self.logger if self.logger is not self.default_logger else None)
self._db_video_path = outputPath
def _addYoutubeMetadata(self):
self._db_yt_title, self._db_yt_description = gpt_yt.generate_title_description_dict(self._db_script)
now = datetime.datetime.now()
date_str = now.strftime("%Y-%m-%d_%H-%M-%S")
newFileName = f"videos/{date_str} - " + \
re.sub(r"[^a-zA-Z0-9 '\n\.]", '', self._db_yt_title)
shutil.move(self._db_video_path, newFileName+".mp4")
with open(newFileName+".txt", "w", encoding="utf-8") as f:
f.write(
f"---Youtube title---\n{self._db_yt_title}\n---Youtube description---\n{self._db_yt_description}")
self._db_video_path = newFileName+".mp4"
self._db_ready_to_upload = True
|