search

MMSB-MOBI / martinize-db

Other
2 stars 0 forks source link

readme

Martini Database Server

Provides JWT-authentification-based REST API for MArtini Database, Molecule Coarse-Grainer and Membrane Builder.

This program serves the MArtini Database server (MAD).

Requirements:

Deploy

An advanced deploy procedure has been written to Deploy.md file. Please ignore this file and see it instead.

Setting up the project

The following procedure assume GROMACS is installed, all your virtual environnements are set properly or your python executables are aware of martinize, insane and other requirements.

First, clone the project with git.

git clone https://github.com/alkihis/martinize-db.git
cd martinize-db

Create JWT keys

This server use JSON Web Tokens, which requires to sign components with RSA keys. If you've cloned the project, you must create the RSA keys with the section below.

mkdir .keys
cd .keys

# do not enter passphrase, press enter
ssh-keygen -t rsa -b 4096 -m PEM -f key_new.pem
openssl rsa -in key_new.pem -pubout -outform PEM -out key_new

Get started

Install all the required dependencies with npm.

npm i

Install TypeScript

Ensure you have TypeScript installed.

tsc -v # 3.8+

You can also use project local TypeScript compiler.

npm run tsc -- -v # 3.8+

Make sure you have rights to create files and folders in project directory, otherwise the server can't work.

Start developement server

tsc  # Compile: will create the build/ folder
npm run start-dev

If everything goes well, it should be configured properly.

Server parameters

The server obey to some parameters in order to run properly. Take a look to src/constants.ts to know more.

You also have CLI args to give to server.

Options:
  -c, --couchdb-url <url>      Couch DB URL (default: process.env.COUCHDB_HOST || constants.URLS.COUCH)
  --server-url <url>           Server URL (default: process.env.SERVER_URL || constants.URLS.SERVER)
  -p, --port <port>            Emit port (default: 4123)
  --wipe-init                  Destroy the CouchDB bases and re-create them.
  --init-db                    Init the required CouchDB databases.
  --quit-after-init            If --wipe-init or --init-db are set, this will quit server after init.
  -l, --log-level <logLevel>   Log level [debug|silly|verbose|info|warn|error] (default: "info")
  --file-log-level <logLevel>  Log level (written to file) [debug|silly|verbose|info|warn|error] (default: "info")
  --log-file <logFile>         File log level

Example:

SERVER_URL=https://martinize-db.ibcp.fr
COUCHDB_HOST=http://localhost:5984

# Run the server on part 4123, with CLI log level verbose, 
# also log in file server.log with minimum level of info
npm run start -- -p 4123 -l verbose --file-log-level info --log-file server.log -c $COUCHDB_HOST --server-url $SERVER_URL

Command line interface

Once started, the server have a CLI to interact with. You can inspect quickly which molecules are inserted, create an user, init a database, send test mails and more. Type help then enter to have informations.

First start

At server first start, you must set up the environnement in order to make user system and database work.

The console will remind you to do it:

Welcome to Martinize server CLI. For help, type "help".

WARN: The database seems to be un-initialized. Please create all databases by entering "database create all".
WARN: Once database is created, you can create an administrator account with "user create".

First, create the databases inside the CLI by typing: database create all.

Then, you must create the first administrator user in order to accept accounts.

> user create
You're about to create a new user.
To exit user creation, type ".exit"
New user name: <enter admin username>
New user email address: <enter desired e-mail address>
New user role. Available roles: "curator" or "admin": admin
New user password: <enter here the desired password>
cli: User has been successfully created.

To be sure the user has been created, use the CLI:

> user get all
cli: [
  {
    _id: '<here is the user ID>',
    _rev: '<here is the couch rev ID>',
    id: '<here is the user ID>',
    name: 'username',
    email: 'user@its.email',
    role: 'admin',
    created_at: '2020-03-09T09:44:02.851Z',
    password: '$2b$10$lJLO7NFdYFVXVEX1VbtcCOH9WnFK35J/1efSFJQSoTM.Dnot7y0pm',
    approved: true
  }
]

You can now login on the website !

Deploy

Compile and prepare deployment

You must have the martinize-db-client project on your computer.

Please see in Deploy.md part Include front-end website to this server to compile and type properly martinize db client.

In this project

This project must have been built at least one time with tsc. Make sure build folder exists.

Compile the required files in a directory, and zip them.


# Copy the required files to a tmp folder
mkdir deploy
cp -R .keys build static templates package.json settings.json deploy/

# Zip the folder
zip -r deploy.zip deploy

# Remove the temporary folder
rm -rf deploy

You now have the required file deploy.zip in order to deploy with Docker! Follow instuctions of Deploy with Docker part.

Deploy with Docker

In order to deploy, docker and docker-compose must be installed and configured properly. Docker is not configured/installed? See Configure_Docker_IBCP.md file.

Create the containers

docker-compose will do the job, following rules defined in the docker-compose.yml file.

docker-compose up -d --build

Containers are ready to be used after docker-compose. Martini Database server default port 4123 will be redirected to host port 5000.

Initialize the server

Two containers are started: martinize_couch_db and martinize_db.

In martinize_db, server is not started automatically, we must initialize the server inside the container.

TODO: inside the container, must install Python3, Python2, insane+martinize2, compile GROMACS.

Enter inside the container, and start the shell script who automatically set a screen with the server.

docker exec -it martinize_db bash  # Open a interactive bash shell in container
# source docker_start_server.sh  # Run the script

Stop or remove the deployment

To stop/remove (f.e., to re-deploy it), remove container and remove its image.

docker-compose stop && docker-compose rm -f

# To reset couch: Remove container volumes
docker-compose rm -v