pyci

pyci is python wrapper to easily deploy multiple instances ShinyProxy apps depending on config declared in json file.

usage

In json can be declared any field from application.yml supported by ShinyProxy (see: https://www.shinyproxy.io/configuration/). Example json:

{
   "ci":[
         {
            "user": "master",
            "shinyproxy":[{
              "port": 8080,
              "display-name": "Some app - prd",
              "port-range-start": 20000,
              "port-range-max": 20100,
              "container-volumes": "['/home/project/data:/data', '/home/project/data/config:/data/config']",
              "container-cmd": "['R', '-e shiny::runApp(\"/data/myapp/inst/shiny\")']"
            }]

         },
         {
            "user": "[email protected]",
            "shinyproxy":[{
              "port": 8080,
              "display-name": "Some app",
              "port-range-start": 20101,
              "port-range-max": 20200,
              "container-volumes": "['/home/project/data:/data', '/home/project/data/config:/data/config']",
              "container-cmd": "['R', '-e shiny::runApp(\"/data/myapp/inst/shiny\")']"
            }]

         }
      ]
}

Example deployment code:

import os
import json
from tempfile import mkstemp
from pyci.yaml import *
from pyci.shinyproxy import *

new_json = filter_json_by_user("[email protected]", "ci/deployment_config.json")

new_json_path = mkstemp()[1]

with open(new_json_path, "w") as f:
    json.dump(new_json, f)

# required variables
jar_name = "shinyproxy.jar"
deploy_cmd="nohup java -jar {0} &".format(jar_name).split()
yaml_path =  "application.yml"

# create custom yml file
# you can replace some application.yml's field on the fly, e.g.
# it can be useful in CI when new docker image is build 
# and you want to deploy app 'ont the fly'  
docker_image_fullname = "some-new-docker-image-to-be-inserted-to-new-yaml"
make_custom_yaml(yaml_path, "container-image", docker_image_fullname, yaml_path)

# before deploying ShinProxy you can test if app successfully runs in container
# id is the id of app in application.yaml
run_docker_cmd_from_yaml(yaml_path, json_path = new_json_path user = "<git email or branch name, e.g. develop">, id = "01_hello")

# deploy ShinyProxy
output = deploy(json_path=new_json_path,
                yaml_path=yaml_path,
                deployment_dir="location-where-to-put-jar-file-and-application-yml",
                deploy_cmd=deploy_cmd,
                url="https://www.shinyproxy.io/downloads/shinyproxy-2.0.5.jar",
                jar_name=jar_name)

deploy.py

There is deploy.py script to easily parametrize your deployment. Example command:

python3 deploy.py --image=<some docker imag> --work-dir=<working dir> --user=develop --config=deployment_config.json --app-id=your_app_id_in_application.yml

To see help run deploy.py -h.