- Set up AWS credentials in AWSCredentials class in aws_helpers.py script.
- Create and place the key file in the ./aws directory (where Boto3 checks, check documentation).
- Upload input.json from ./data to the S3 bucket named 'ytdlinput'.
- Install requirements from requirements.txt
- Create and add ffmpeg in the ./dep directory (./dep/ffmpeg/bin/ffmpeg.exe).
- Upload input.json from ./data to the S3 bucket named 'ytdlinput'.
- Set RUN_LOCALLY=True and USE_RAY to True or False if you want to run it using Ray or sequentially.
- For Windows, in WSL CLI run:
- Create the cluster:
ray up -y /mnt/c/Users/Stefan/Desktop/Personal/Projects/Soniox/yt_dl_project/ray/yt_dl_cluster.yaml
- Forward ports locally:
ray dashboard /mnt/c/Users/Stefan/Desktop/Personal/Projects/Soniox/yt_dl_project/ray/yt_dl_cluster.yaml
- Submit task:
RAY_ADDRESS='http://localhost:8265' ray job submit --runtime-env-json='{"working_dir": "./", "pip": ["boto3", "boto3-stubs[dynamodb]", "scrapetube", "youtube-transcript-api", "youtube-wpm"]}' -- python ./pipeline.py
- Create the cluster: