CLI

This section provides detailed information on how to use the GRID Enterprise Command-Line Interface (CLI). The GRID CLI acts as an orchestration module for GRID sessions. Run it by typing grid at the terminal in the appropriate Python virtual environment. The CLI manages GRID sessions and Docker containers, and it depends on configuration files.

When initializing the GRID CLI for the first time, you will be prompted to provide credentials for the resource configuration file, which are then stored in ~/.grid/resource_config.json. The credentials will be provided to you during your onboarding process for the GRID Enterprise platform.

Along with the credentials, the resource configuration file also contains information about the nodes where you wish to run GRID sessions, as well as information about storage for saving GRID session data. For your convenience, the resource configuration file targets localhost by default, which allows you to run sessions on the same machine. If you wish to target other remote machines, see the following sections.

Configuration Overview

Machine Configuration: The ~/.grid/resource_config.json file allows you to define remote machines. Note that the "local" section is present by default. To use a remote machine, ensure that you add the necessary keys for IP address, username, and password. For example:
```
{
  "resources": {
    "local": {},
    "remote": {
      "ip": "192.168.0.106",
      "username": "usernamegoeshere",
      "password": "passwordgoeshere"
    }
  }
}
```
When running GRID commands, always verify your node names (e.g., @local, @node0) before executing to ensure you are running them on the right machines.
Mounting Storage: To mount a folder from the host to the GRID containers, add a storage key under the corresponding resource. For example, if you wish to mount a folder from localhost into the GRID containers:
```
{
  "resources": {
    "local": {
      "storage": {
        "code": "/home/test/Documents/code"
      }
    }
  }
}
```
This will mount /home/test/Documents/code on the host to /mnt/code in the GRID containers.
Ports for Remote Access: If you are using GRID on a remote machine, ensure that the following ports are open to access the simulation, notebooks, and visualizations:
- 8000: GRID server
- 8890: Jupyter Notebook
- 3080: Simulation
- 9090: Rerun visualization
- 9877: Rerun stream

Commands Overview

Below is a list of the available commands in the GRID Enterprise CLI. Resource targeting is done using the @resource_name syntax, defaulting to @local if no resource is specified.

The first step is to log into the GRID platform and initialize the containers. Open your terminal and run the following command:

grid

login

Logs into the GRID registry using the username and password or access token. The resource can be specified using @resource_name.

login @local

init

Spins up the GRID containers on the specified node (defaults to local` if no node is specified).

init @local

terminate

Terminates the GRID containers on the specified node.

terminate @local

update

Updates the GRID containers on the specified node. You must be authenticated to the container registry for this to work.

update @local

session start <session_id> <config_path> @<resource_name>

Starts a session with the specified session ID and configuration path on the chosen resource (defaults to @local` if no resource is specified).

session start test_session ./config/session_config.json @node0

session stop <session_id>

Stops the session with the specified ID.

session stop test_session

session list

Lists all ongoing sessions.

session list

open <session_id> nb | sim | viz

Opens the corresponding entity for the specified session:

nb: Jupyter notebook
sim: Simulation stream
viz: Rerun visualization

open test_session nb
open test_session sim
open test_session viz

help

Displays a list of all available commands and their functions.

help