Skip to content

Latest commit

 

History

History
35 lines (32 loc) · 2.27 KB

README.md

File metadata and controls

35 lines (32 loc) · 2.27 KB
Raw

Webhook configuration and scripts

This folder documents the server settings needed to host JPhyloRef as a web server using Webhook. This requires the webhook binary to be downloaded to the same folder as these scripts.

Webhook will set up a web server that listens for HTTP requests on the specified port and will execute a program for each request that arrives. If you would like your web server to run on that webserver, you can have webhook run JPhyloRef directly with the "--errors-as-json" flag. If you need your web server to run JPhyloRef on a cluster that uses SLURM as its job manager, you can use the included exec_jphyloref_webhook.sh script to start the job instead.

The files in this folder are:

  • start.sh should be executed to start the server. It sets up the Access-Control-Allow-Origin and Access-Control-Allow-Headers HTTP headers to ensure that requests from other websites via CORS are allowed. The $PORT environment variable should be set to the HTTP port to listen to.
  • hooks.json is the settings file for Webhook. It expects a jsonld form value containing a JSON-LD document to be reasoned over. It uses X-Hub-Signature to ensure that the browser knows a secret (which should be set as $SECRET) shared with the reasoner.
  • exec_jphyloref_webhook.sh is a helper script that executes JPhyloRef and prints the JSON document containing resolved phyloreferences to standard output. It calls JPhyloRef with particular settings at three levels:
    • Some settings (number of tasks, number of CPUs per task, timeout) are sent to srun, which starts JPhyloRef on a cluster.
    • Some settings (maximum memory, HTTP proxy settings) are sent to Java to set up the environment in which JPhyloRef can run.
    • Some settings (the name of the input file) are sent to JPhyloRef for processing.