init commit

.env.example

@@ -0,0 +1,4 @@
+CLIENT_SECRETS_FILE=src/client_secrets.json
+SCOPES=https://www.googleapis.com/auth/youtube.upload
+API_SERVICE_NAME=youtube
+API_VERSION=v3

.gitignore

@@ -0,0 +1,8 @@
+venv/*
+temp/*
+output/*
+**/client_secrets.json
+.pytest_cache/*
+**/__pycache__/*
+.DS_Store
+.env

LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2024 FujiwaraChoki
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

README.md

@@ -0,0 +1,58 @@
+# Youtube-short creater with BERT-VITS2 & DALLE-3 & ChatGPT 🚀
+
+Automate the creation of YouTube Shorts locally, simply by providing a video topic to talk about.
+
+[![ビデオの例](docs/thumbnail.png)](https://www.youtube.com/shorts/Eah3Pb9s2Kk "例を見る")
+
+
+## Installation 📥
+
+`YoutubeShortCreator` requires Python 3.11 to run effectively. If you don't have Python installed, you can download it from [here](https://www.python.org/downloads/).
+
+After you finished installing Python, you can install `YoutubeShortCreator` by following the steps below:
+
+```bash
+git clone https://github.com/avonx/YoutubeShortCreator.git
+cd YoutubeShortCreator
+
+# Install requirements
+pip install -r requirements.txt
+
+# Copy .env.example and fill out values
+cp .env.example .env
+
+# Run the python program
+python generate_video.py
+
+# Alternatively, run this to upload videos to youtube automatically.
+python upload_video.py
+```
+
+See [`.env.example`](.env.example) for the required environment variables.
+
+## Usage 🛠️
+
+1. Copy the `.env.example` file to `.env` and fill in the required values
+2. Run this command `python generate_video.py`
+3. Enter a topic to talk about
+4. Wait for the video to be generated
+5. The video's location is `output/<タイトルは自動的に生成されます>.mp4`
+
+
+## Automatic YouTube Uploading 🎥
+
+To use this feature, you need to:
+
+1. Create a project inside your Google Cloud Platform -> [GCP](https://console.cloud.google.com/).
+2. Obtain `client_secret.json` from the project and add it to the Backend/ directory.
+3. Enable the YouTube v3 API in your project -> [GCP-API-Library](https://console.cloud.google.com/apis/library/youtube.googleapis.com)
+4. Create an `OAuth consent screen` and add yourself (the account of your YouTube channel) to the testers.
+5. Enable the following scopes in the `OAuth consent screen` for your project:
+
+```
+'https://www.googleapis.com/auth/youtube.upload'
+```
+
+## License 📝
+
+See [`LICENSE`](LICENSE) file for more information.

requirements.txt

@@ -0,0 +1,12 @@
+termcolor
+openai
+pytest
+black
+pydub
+requests
+google-api-python-client
+google-auth
+google-auth-httplib2
+google-auth-oauthlib
+httplib2
+oauthlib

src/generate_ideas.py

@@ -0,0 +1,59 @@
+from openai import OpenAI
+import json
+from termcolor import colored
+
+client = OpenAI()
+
+
+def generate_ideas(meta_topic: str, num_ideas: int) -> dict:
+ """
+ メタトピックについて、アイデアを生成する。
+ """
+ # プロンプトの構築
+ prompt_valuables = f"""
+ 下記のメタトピックに応じて、youtubeのショート動画のアイデアを考えてください。
+
+ 主題: {meta_topic}
+ 出力するアイデアの個数: {num_ideas}
+ """
+
+ prompt_format = """
+ 次の形式のjsonで出力するようにしてください。
+ "ideas": [
+ "アイデア1", "アイデア2", ...
+ ]
+
+ 例）
+ "ideas": [
+ "実は宇宙人はすでに紛れ込んでいる", "ドラえもんが怒った時に何が起こるか", ...
+ ]
+ """
+ response = client.chat.completions.create(
+ model="gpt-4-1106-preview",
+ # model="gpt-3.5-turbo-1106",
+ messages=[
+ {"role": "system", "content": "You are a helpful assistant."},
+ {
+ "role": "user",
+ "content": f"次の文章からJSONを生成してください。\n{prompt_valuables}{prompt_format}",
+ },
+ ],
+ response_format={"type": "json_object"},
+ )
+
+ script = response.choices[0].message.content.strip()
+ # scriptをjsonに変換
+ try:
+ script = json.loads(script)
+ return script
+ except json.JSONDecodeError as e:
+ print(colored("スクリプトの生成に失敗しました。", "red"))
+ print(colored("エラー内容: ", "red"), e)
+ return None
+
+
+if __name__ == "__main__":
+ meta_topic = "SFチックな、ちょっとオカルトな話を具体的に。"
+ num_ideas = 2
+ ideas = generate_ideas(meta_topic, num_ideas)
+ print(ideas)

src/generate_video.py

@@ -0,0 +1,52 @@
+from pathlib import Path
+from termcolor import colored
+
+from generate_ideas import generate_ideas
+from topic2text import generate_script
+from text2voice import generate_audio_for_clips
+from voice2video import process_json_data, create_combined_video_from_clips
+
+
+def topic2video(video_subject, num_clips, output_file_path):
+ # Create or clear the temp folder
+ temp_folder = Path("./temp")
+ temp_folder.mkdir(exist_ok=True)
+ for file in temp_folder.glob("*"):
+ file.unlink()
+
+ clips = generate_script(video_subject, num_clips)
+ print(colored("[+] Generating script...", "yellow"))
+ print(clips)
+
+ updated_clips = generate_audio_for_clips(clips)
+ print(colored("[+] Generating audio for clips...", "yellow"))
+ print(updated_clips)
+
+ updated_clips = process_json_data(updated_clips)
+ print(colored("[+] Processing JSON data...", "yellow"))
+ print(updated_clips)
+
+ video_path = create_combined_video_from_clips(
+ updated_clips, output_file_path)
+ print(colored("[+] Creating combined video...", "yellow"))
+
+ # Delete temporary files
+ for file in temp_folder.glob("*"):
+ file.unlink()
+
+ return video_path, updated_clips
+
+
+if __name__ == "__main__":
+ meta_topic = "科学に関する話題を具体的に。"
+ num_ideas = 3
+ ideas = generate_ideas(meta_topic, num_ideas)
+ print(colored("[+] Generating ideas...", "yellow"))
+ print(ideas)
+
+ for idea in ideas["ideas"]:
+ video_subject = idea
+ num_clips = 5
+ output_file_path = f"./output/{video_subject}.mp4"
+ video_path, updated_clips = topic2video(
+ video_subject, num_clips, output_file_path)

src/text2voice.py

@@ -0,0 +1,115 @@
+import requests
+from pydub import AudioSegment
+import io
+import uuid
+from pathlib import Path
+from termcolor import colored
+
+# tempフォルダのパスを定義（存在しない場合は作成）
+temp_folder = Path("./temp")
+temp_folder.mkdir(exist_ok=True)
+
+
+def generate_audio_for_clips(clips):
+ """
+ Generate audio files for each clip in the given list.
+
+ Args:
+ clips (list): A list of clips, where each clip is a dictionary containing the "text" key.
+
+ Returns:
+ None
+ """
+ for clip in clips["clips"]:
+ filename = f"{str(uuid.uuid4())}.mp3"
+ filepath = temp_folder / filename # 保存するファイルパス
+ text = clip["text"] # 音声に変換するテキスト
+
+ combined_audio = process_and_combine_audio(text)
+ if combined_audio:
+ combined_audio.export(filepath, format="mp3")
+ clip["audio_path"] = str(filepath) # 生成した音声ファイルのパスをJSONに追加
+ else:
+ clip["audio_path"] = None # 音声生成に失敗した場合
+ return clips
+
+
+def process_and_combine_audio(text: str) -> AudioSegment:
+ """
+ Process and combine audio segments for each paragraph in the given text.
+
+ Args:
+ text (str): The input text.
+
+ Returns:
+ AudioSegment: The combined audio segment.
+
+ """
+ text_no_newlines = text.replace("\n\n", "")
+ paragraphs = [f"{p}。" for p in text_no_newlines.split("。") if p]
+
+ combined_audio = None
+ for i, paragraph in enumerate(paragraphs):
+ print(
+ colored(f"[+] Processing paragraph {i+1}/{len(paragraphs)}...", "yellow"))
+ audio_content = get_tts_audio(paragraph)
+ if audio_content:
+ current_audio = AudioSegment.from_file(
+ io.BytesIO(audio_content), format="wav")
+ silence = AudioSegment.silent(duration=1000) # 1秒の無音を追加
+ combined_audio = combined_audio + silence + \
+ current_audio if combined_audio else current_audio
+ return combined_audio
+
+
+def get_tts_audio(text: str) -> bytes:
+ """
+ Get text-to-speech audio for the given text.
+
+ Args:
+ text (str): The text to convert to audio.
+
+ Returns:
+ bytes: The audio content in bytes if successful, None otherwise.
+ """
+ url = "https://igeb37tgif2guh-5000.proxy.runpod.net/voice"
+ params = {
+ "text": text,
+ "encoding": "utf-8",
+ "model_id": 4,
+ "speaker_id": 0,
+ "sdp_ratio": 0.6,
+ "noise": 0.7,
+ "noisew": 1.1,
+ "length": 0.75,
+ "language": "JP",
+ "auto_split": True,
+ "split_interval": 0.5,
+ "assist_text_weight": 1,
+ "style": "Neutral",
+ "style_weight": 5,
+ }
+ response = requests.get(url, params=params)
+ if response.status_code == 200 and "audio/wav" in response.headers.get("Content-Type"):
+ return response.content
+ return None
+
+
+if __name__ == "__main__":
+ # JSONデータを処理するための例
+ clips = {
+ "clips": [
+ {
+ "num": 0,
+ "title": "サンプルタイトル",
+ "text": "これはサンプルのテキストです。宇宙人はいます。会ったことがあります！",
+ "video_prompt": "静かな森の風景を描写する文章。",
+ },
+ # 他のクリップ...
+ ]
+ }
+
+ updated_clips = generate_audio_for_clips(clips)
+
+ # 結果を確認
+ print(updated_clips)