Skip to content

Latest commit

 

History

History
69 lines (48 loc) · 1.76 KB

README.md

File metadata and controls

69 lines (48 loc) · 1.76 KB
Raw

bmsurge-renderer/worker

Environment variables configuration

# .env
GOOGLE_APPLICATION_CREDENTIALS=/data/private/<service-account>.json
BMSURGE_RENDER_OUTPUT_BUCKET=<gcs-bucket-name>
BMSURGE_WORKER_GCR_REPOSITORY=gcr.io/<gcp-project>/<repo-name>

Build Docker image

./scripts/build

Try out locally

The ./script/run shell script will mount the src directory into the container, so that you don't need to rebuild the container everytime.

Rendering process

./scripts/run render https://.../yamajet_is_sugoi.zip -o /data/output/foon.mp3

This will download the package and renders an MP3 file. After running this, you will find the MP3 file at ./tmp/output/foon.mp3. Metadata about the rendering process is printed out to the console.

Server

To test the server locally, you need to put a service account file into ./private directory and reference it inside .env file using the GOOGLE_APPLICATION_CREDENTIALS environment variable.

The service account used needs to have access to the Google Cloud Storage bucket named by the BMSURGE_RENDER_OUTPUT_BUCKET environment variable.

./scripts/run server

This will start a server on port 4567. Invoke it using httpie:

http -S --timeout=900 put http://localhost:4567/renders/00000000-0000-0000-0000-000000000000 url='https://.../yamajet_is_sugoi.zip'

The MP3 file shall appear at gs://<bucket>/00000000-0000-0000-0000-000000000000.mp3

Build and push Docker image to GCR

./scripts/push

Deployment

See: https://cloud.google.com/run/docs/deploying

  • Set memory limit to 2048 MB.
  • Set request timeout to 900 seconds.
  • Required environment variables:
    • BMSURGE_RENDER_OUTPUT_BUCKET
  • Allow the service account to access the Google Cloud Storage bucket.