living-meta-analysis

A tool for supporting living and collaborative meta-analysis.

pre-requisites

node.js and npm (version TBS)

installation and running

clone the repo
npm install
fill in google-datastore-specific settings in server/config.js

the above needs an authentication key pointed at by config.gcloudProject.keyFilename – this is generated for you by the Google developer console

change webpages/js/auth.js to include your own client ID also generated by Google
fill in HTTPS settings, or comment out HTTPS-specific parts of config.js to run on HTTP only
npm start

production server updating

ssh lima
sudo login -f lima
cd living-meta-analysis
# do not forever stop server
git stash
git pull
git stash apply
git diff -U2             # review everything
npm install              # if package.json has changed
forever restart server   # if server has changed
logout                   # so we don't have a hanging terminal to the server

invite codes

For now with ./invites.txt to generate an invite, append a line like this:

random-code-12345 # 2017-02-28 generated for Cochrane workshop attendees

use the script geninvite to generate invites, e.g.

for pom in `seq 1 100`; do geninvite "for Cochrane workshop attendees and other invitees"; done >> invites.txt

Then print them with LiMA's print page.

Source of geninvite:

#!/bin/bash

code=''
while grep -q "$code" invites.txt
do
  code=`uuidgen`
  code=${code##*-}
done

echo $code '#' `date +"%Y-%m-%d %H:%M"` $1

database dumps and restores

We have server-side datastore dump and restore scripts. They communicate directly with the datastore as configured in config.js; the LiMA server should be down while the restore script is running.

To dump the currently configured data store:

npm run db-dump > dumpfile

To restore from a dumpfile, first stop LiMA server, then run the following:

npm run db-add < dumpfile

After restoring data, start LiMA server again and it should see the new data.

Jnnybrrn / living-meta-analysis