-
Notifications
You must be signed in to change notification settings - Fork 5
Installation Option A: Pull a Prebuilt Image (Recommended)
robinpaul85 edited this page May 22, 2024
·
20 revisions
This may be the easiest option to get started. However, it may require more Docker know-how. for customization.
Make sure, the docker client is running.
TAG=latest # can change to a version number like 2.11.2
IMAGE_NAME=ghcr.io/stjude/ppfull:$TAG # may use ppserver:$TAG for server-only image
docker pull $IMAGE_NAME
# download the run helper scripts
wget https://raw.githubusercontent.com/stjude/proteinpaint/master/container/run.sh
wget https://raw.githubusercontent.com/stjude/proteinpaint/master/container/createPPNetwork.sh
wget https://raw.githubusercontent.com/stjude/proteinpaint/master/container/verify.sh
wget https://raw.githubusercontent.com/stjude/proteinpaint/master/container/validateConfig.js
wget https://raw.githubusercontent.com/stjude/proteinpaint/master/utils/public-support-files-download.sh
chmod a+x *
Download support files (such as appropriate reference genome files) using the command ./public-support-files-download.sh
.
Test serverconfig.json
{
"debugmode": true,
"defaultgenome": "hg38",
"genomes": [
{
"name": "hg19",
"species": "human",
"file": "./genome/hg19.js",
"datasets": []
},
{
"name": "hg38",
"species": "human",
"file": "./genome/hg38.js",
"datasets": []
}
],
"tpmasterdir": "/path/to/tp_dir",
"cachedir": "/path/to/cache_dir",
"URL": "http://localhost:3456",
"gfClient": "/home/root/pp/tools/gfClient",
"port": 3000,
"backend_only": false
}
To test, make sure that your current working directory has
- a serverconfig.json which has
- a
tpmasterdir
entry, for the absolute path to the data directory - a
"URL": "http://localhost:[PORT]"
entry (default PORT=3456, can be set to any valid, non-conflicting numeric port value)
- a
- an optional dataset folder, containing js files of any serverconfig.genomes.datasets[] entryvthat is not already included in proteinpaint/server/dataset
./run.sh $IMAGE_NAME
# open the browser to your serverconfig.URL entry
# example routes to check, assuming serverconfig.URL=http://localhost:3456
# http://localhost:3456/healthcheck
# http://localhost:3456/genomes
# http://localhost:3456 should open the Proteinpaint landing page
# Hints:
# - inspect logs with 'docker logs pp'
# - ssh into the container with 'docker exec -it pp'
# - stop the container with 'docker stop pp'