FlowServ Demo App Using Streamlit.io

About

Depreciation: The flowserv-app code has been integrated in flowserv-core starting with version 0.9.0. This repository is therefore no longer maintained.

This package allows to run flowServ workflow templates as streamlit.io applications.

Installation

The app currently has to be installed from GitHub directly. It is recommended that you use a virtual environment for all the following commands.

Install the flowapp package:

pip install git+https://github.com/scailfin/flowserv-app.git

To configure the app make sure to set the environment variables FLOWSERV_BACKEND_CLASS and FLOWSERV_BACKEND_MODULE. These environment variables determine the backend that is used to run workflows. See the flowServ documentation for more details.

The repository contains a default configuration file env.config. You can use this default configuration (e.g., by running source env.config) to define the default multi-process workflow engine as the backend.

Running the Application

The command to run a flowServ workflow as a streamlit web application is:

streamlit run flowapp/app.py [ --
    [-w | --source=] <workflow-identifier or workflow-template-folder>
    [-s | --specfile=] <filename>  # Optional path to workflow specification file.
    [-m | --manifest=] <filename>  # Optional path to workflow manifest file.
    [-n | --name=] <name>          # Optional application title.
]

The __source references the workflow that is being run. This can either be the identifier of a template from the flowServ Workflow Repository or the path to a local folder containing a workflow template.

As an example, to run the Hello World Demo use the following command:

streamlit run flowapp/app.py -- --source=helloworld