Skip to content

Latest commit

 

History

History
84 lines (54 loc) · 2.48 KB

INSTALLATION.md

File metadata and controls

84 lines (54 loc) · 2.48 KB
Raw

Installation & Deployment

Requirements

  • node >= v10
  • npm >= v6

Usage

  1. Install packages with: npm install

  2. Run with: npm run dev

Configuration

When running Gecko in server mode, you'll be able to configure the host port, create a .env file with:

GECKO_SERVER_CONTAINER_PORT=<<<host port>>

AWS integration

S3 storage

You can use a S3 storage account to:

  • Load audio & annotation from a bucket
  • Save annotations to a bucket

You need to add these environment variables to your .env file:

AWS_BUCKET=<<<put your aws bucket name here>>>
AWS_REGION=<<<put the region where your bucket is, something like us-east-3>>>
AWS_ACCESS_KEY_ID=<<<aws access key id>>>
AWS_SECRET_ACCESS_KEY=<<<aws secret access key>>>
AWS_FOLDER=<<<the folder containing your files (audio & annotations)>>>

At this point, you can load data from S3 Storage using this url:

http://localhost:8080/?save_mode=server&audio=http://localhost:8080/s3_files/file.mp3&json=http://localhost:8080/s3_files/annotation.json

Notes:

  1. Gecko will aggregate the AWS_FOLDER with the file name, do not use the object key S3 gives you, only the file name is required.
  2. You can use any annotation format available in Gecko, just replace the key of the parameter, for exemple with rttm : http://localhost:8080/?save_mode=server&audio=http://localhost:8080/s3_files/file.mp3&rttm=http://localhost:8080/s3_files/annotation.rttm
  3. On save, Gecko will override the annotation file and save another file with a timestamp, to maintain some kind of primitive history mechanism.
  4. You can have multiple annotation files, just aggregate them using ; : http://localhost:8080/?save_mode=server&audio=http://localhost:8080/s3_files/file.mp3&json=http://localhost:8080/s3_files/annotation_1.json;http://localhost:8080/s3_files/annotation_2.json

Example:

For an audio file and corresponding annotation file: s3://example-audio-bucket/my-audio/file.mp3 and s3://example-audio-bucket/my-audio/annotation.json the .env file would be:

AWS_BUCKET=example-audio-bucket
AWS_REGION=<<<something like us-east-3>>>
AWS_ACCESS_KEY_ID=<<<aws access key id>>>
AWS_SECRET_ACCESS_KEY=<<<aws secret access key>>>
AWS_FOLDER=my-audio # no leading or trailing "/"

Deploy

Client mode

npm run build

Copy the build folder to an annotator machine.

Run index.html

Server mode

npm run build

npm run server

Annotator can now access the server remotly, and can use S3 integration if configured.