seamm-dashboard 2022.9.29

The Web Dashboard for SEAMM (Simulation Environment for Atomistic and Molecular Simulations).

MolSSI Dashboard (Flask-coreUI)

This application is a results Dashboard for MolSSI projects.

The server runs Flask on Passenger and Apache, or can run flask testing server.

Install the SEAMM dashboard

Install from PyPi

You can install the SEAMM Dashboard from PyPi. First create a conda environment for the dashboard.

conda create -n seamm-dashboard python=3.9
conda activate seamm-dashboard

Next, install from conda

pip install seamm-dashboard

Install from this repository

To get the most up-to-date version of the dashboard, install the package from this repository.

git clone https://github.com/molssi-seamm/seamm_dashboard.git
cd seamm_dashboard

To create the seamm-dashboard environment and install the necessary packages, type

$ make environment

in the top level of your directory.

After the script is finished running, activate the SEAMM Dashboard conda environment:

conda activate seamm-dashboard

Installing the Datastore (Required)

You must also install the SEAMM Datastore. This is the package that manages the database connection. We are working on getting it on PyPi. For now, navigate to the repository. You should clone this repository and install the package. Make sure you are not in the SEAMM dashboard directory if you installed from GitHub in the previous step.

git clone https://github.com/molssi-seamm/seamm_datastore.git
cd seamm_datstore
pip install .

If your conda environment is activated, you're ready to start running the dashboard.

Running the dashboard

You can then run the dashboard after you have installed and activated the dashboard environment.

Run a demo dashboard

If you do not have SEAMM installed, you can view a demo dashboard by using the data in this repository. Use the command

./results_dashboard.py --initialize --datastore $(pwd)/data --jwt-secret-key 'super-secret' --secret-key 'another-super-secret'

If you are running the dashboard in production, you should use better secrets.

Open a browser and navigate to http://localhost:5000/ to view the sample dashboard. Running the sample dashboard will create a user in the database with the same username you use on your computer the default password is default.

Running with SEAMM installed

If you have SEAMM installed, you can connect to your seamm datastore. In the top level of the repository, type the following command into the terminal:

results_dashboard.py

The dashboard can then be viewed in your browser at localhost:5000. By default, the dashboard uses the location of the datastore in ~/.seamm/seamm.ini to locate the datastore to display. This can, however, be overridden by a command line argument --datastore xxxx. There are other options available. For more information run

results_dashboard.py --help

usage: results_dashboard.py [-h] [--dashboard-configfile DASHBOARD_CONFIGFILE] [--datastore DATASTORE] [--initialize] [--no-check]
                            [--log-level {CRITICAL,ERROR,WARNING,INFO,DEBUG,NOTSET}] [--console-log-level {CRITICAL,ERROR,WARNING,INFO,DEBUG,NOTSET}] [--log_dir LOG_DIR]
...
optional arguments:
  -h, --help            show this help message and exit
  --dashboard-configfile DASHBOARD_CONFIGFILE
                        a configuration file to override others (default: None)
  --datastore DATASTORE
                        The datastore (directory). [env var: SEAMM_DATASTORE] (default: .)
  --initialize          initialize, or reinitialize, from the job files [env var: INITIALIZE] (default: False)
  --no-check            do not check that jobs are in the database [env var: NO_CHECK] (default: False)
  --log-level {CRITICAL,ERROR,WARNING,INFO,DEBUG,NOTSET}
                        the logging level for the dashboard [env var: LOG_LEVEL] (default: INFO)
...

By default, if the database does not exist, it will be initialized from the job files in the datastore. Otherwise, the dashboard will scan the job files on startup and add any missing ones to the database. You can prevent this initial check with --nocheck. Similarly, if you wish to force the database to be recreated from scratch, use the --initialize flag.