Skip to content

Latest commit

 

History

History
280 lines (186 loc) · 6.06 KB

productionDeployment.md

File metadata and controls

280 lines (186 loc) · 6.06 KB
Raw

BioCompute Portal Local Deployment Instructions/notes

System Setup

Requirements

  • Node.js
  • Python 3
  • PyEnv (optional but reccomended)

BCO_API

The BCO API repository contains a top-level folder “admin_only” which contains service definitions for gunicorn and django. Thus, we can first clone the repository (only necessary to clone once since we can use the shell scripts in the repository after the initial cloning to administer the application), then we can perform a few administrative steps to start the API.

Clone the repository

git clone https://github.com/biocompute-objects/bco_api

cd bco_api

Make sure you are on the desired branch (Check for latest branch):

git switch 22.05

Enter the repository, create a virtual environment, and install the required packages

pyenv local 3.9.4

python3.9 -m venv env

source env/bin/activate

python -m pip install -r requirements.txt

(You can use python3.9 if you’d like, but since you’re in the virtual environment you created python points to python3.9)

Modify the Config files:

Check/Edit the server.conf file

This is the main server configuration file for the BCO API. (most of these values will NOT need to be changed for local deployment)

vim bco_api/bco_api/server.conf

Production and publishing flags NOTE: Valid values are True or False (note the capitalization).

[PRODUCTION]
production=False

DB Version

[VERSION]
version=22.01

Is this a publish-only server?

[PUBLISHONLY]
publishonly=False

Security settings: Create a key for an anonymous public user.

[KEYS]
anon=627626823549f787c3ec763ff687169206626149

Which host names do you want to associate with the server? Note: the local hostname (i.e. 127.0.0.1) should come at the end.

[HOSTNAMES]
prod_names=test.portal.biochemistry.gwu.edu,127.0.0.1
names=127.0.0.1:8000,127.0.0.1

Give the human-readable hostnames

[HRHOSTNAME]
hrnames=BCO Server (Default)

The public hostname of the server (i.e. the one to make requests to)

[PUBLICHOSTNAME]
prod_name=https://test.portal.biochemistry.gwu.edu
name=http://127.0.0.1:8000

Who gets to make requests?

[REQUESTS_FROM]
portal=https://test.portal.biochemistry.gwu.edu
local_development_portal=http://127.0.0.1:3000,http://localhost:3000
public=true

Namings: How do you want to name your objects?

[OBJECT_NAMING]
prod_root_uri=https://test.portal.biochemistry.gwu.edu
root_uri=http://127.0.0.1:8000
prod_uri_regex=root_uri/prefix_(\d+)/(\d+).(\d+)
uri_regex=root_uri/prefix_(\d+)/(\d+).(\d+)

**Requests ** Where are the request templates defined?

[REQUESTS]
folder=../api/request_definitions/

Where are the validation templates defined?

[VALIDATIONS]
folder=../api/validation_definitions/

Set up DB

cd bco_api/bco_api

Option #1: Use existing DB

Copy the dev db

cp ../admin_only/db.sqlite3.dev db.sqlite3

superusername: bco_api_user
password: testing123

Make Migrations

python3.9 manage.py migrate

Option #2: Create a new DB

Make Migrations

python3.9 manage.py migrate

Create a super user for the API:

python3.9 manage.py createsuperuser

Then follow the prompts

Then: Do a quick check to make sure the server can run

Start the serer

python3.9 manage.py runserver 8000

Make sure API is accessible via web browser.

[YOU WEB HOST HERE]/api/admin/

EX: http://localhost:8000/api/admin/

If it worked you should be able to login using the SuperUser credentials you created above

BCO Portal Server deployment (portal_userdb/server)

  • For HTTPS access:

git clone https://github.com/biocompute-objects/portal_userdb

  • For SSH access*(RECCOMENDED)*:

git@github.com:biocompute-objects/portal_userdb.git

Then

cd portal_userdb/

Make sure you are on the desired branch:

git switch <whatever branch you need> (or whatever branch you need)

Enter the repository, create a virtual environment, and install the required packages

cd client

pyenv local 3.9.4 (optional)

python3.9 -m venv env

source env/bin/activate

pip3.9 install -r requirements.txt

Modify the Config file

vim portaluserdb/settings.py

Allow all hosts?

29:    ALLOWED_HOSTS=['*']
152:  CORS_ORIGIN_ALL_ALL = True

Generate the secrets file

  • Copy the .secrets.example to .secrets

  • On linux (or MAC) generate a 32-bytes long PSK key using the openssl command for the DJANO_KEY:

openssl rand -base64 32

  • Update the .secrets file with the required keys:
[GOOGLE_KEYS]
DJANGO_GOOGLE_OAUTH2_CLIENT_ID=<your-client-id-here>
DJANGO_GOOGLE_OAUTH2_CLIENT_SECRET=<your-client-secret-here>

[DJANGO_KEY]
SECRET_KEY=<your-Django-secret-key--here>

Set up DB

Option #1: Use existing DB

cp admin/db.sqlite3 .

python3.9 manage.py migrate

superusername: bco_api_user
password: testing123
Option #2: Create a new DB

Create a super user for the API:

python3.9 manage.py migrate

python3.9 manage.py createsuperuser

Run Server

python3.9 manage.py runserver 8080

Make sure API is accessible via web browser. EX:

https://localhost:8080/users/admin/

If it worked you should be able to login using the SuperUser credentials you created above

[YOU WEB HOST HERE]/users/admin/

BCO Portal Client deployment (portal_userdb/client)

cd /client/

Install packages

npm install

Set up the environment file

cp .env.example .env

Then you will have to add values for the following:

REACT_APP_USERDB_URL=http://localhost:8080/users/
REACT_APP_BCOAPI_URL=http://127.0.0.1:8000/api/
REACT_APP_GOOGLE_CLIENT_ID=404409424650-a1hh4j6m9r3998v16siia2lum9un21ip.apps.googleusercontent.com

Start service

npm run start

Log in with the superuser credentials you created or imported

If you copied over the existing dbs you should be able to log in with any of the crednetials listed in /portal_userdb/server/admin/users.tsv

Otherwise you will have to register a new user.