Roadie-GCP

A helper container to execute a program on Google Cloud Platform.

Features:
  • downloads necessary files from web and Google Cloud Storage,
  • runs commands,
  • uploads results to Google Cloud Storage.

Roadie-GCP takes one instruction file. The instruction file is simple YAML documents like

apt:
- nodejs
source: https://github.com/itslab-kyushu/youtube-comment-scraper.git
data:
- http://sample.com/run.sh
- gs://a-project/input/data:/tmp
run:
- npm install
- run.sh /tmp/data
result: gs://a-project/result/
upload:
- "*.out"

This example commands Roadie-GCP to install nodejs via apt, and download source codes from a Github repository. Then, it prepares to data from some web server and Google Cloud Storage, and run a command run.sh /tmp/data. Finally, it uploads stdout and results which have extension .out to a bucket in Google Cloud Storage. Roadie-GCP automatically shutdowns the virtual machine Roadie-GCP is running on, so you can minimize charge.

Indices and tables

License

This software is released under the MIT License.