This is an application that is designed to show off using Python tools to query a database, create a visualization using Altair, and using streamlit to develop a website to perform these actions.

We have deployed on Render.com (it is here: https://wells-202306-vr3d.onrender.com/).

Important notes

We need to provide Render.com with an environmental variable in the configuration called WELLS_URL that has the connection string for connecting with the online database (remember that for security reasons we don't want to include it in the git repository as it has a username and password in that connection string). It is possible to supply one or more environmental variables through the web interface when we create the application.
We should tell Render.com what version of Python we are using, i.e. we developed the application on Python 3.10.8 so we know it should work with that Python version (and, of course, the packages we will install in the environment using the requirements.txt file). Render.com will otherwise default to some Python version of its own choosing.
Render.com is generally assuming we are running a Flask application, but we are using streamlit. We must provide the "startup" command so that Render will correctly start the application. Also, we need to tell Render to use an open port to communicate with the "outside world" and to not attempt to open a web browser on the deployment machine when starting the streamlit application (not only would that be not helpful, I think it might actually cause an error). The way to do this on Render is to use the startup command streamlit run --server.port=$PORT --server.headless=true app.py (replace app.py with whatever the name of your application might be).

Remark

The requirements file has more packages than are necessary for the deployment to Render.com. For example, the ipykernel package was installed only for the purposes of developing the Altair visualization. (Installing that package triggered the installation of several other packages.) python-dotenv was also only installed since we were developing the visualization in a Jupyter notebook, and didn't want to inject the PostGreSQL connection string into the notebook, but were reading it through an environmental variable.
Sometimes we might use two virtual environments, one for the purposes of development, and a second one for the purposes of deployment. This will help keep the deployment environment lightweight, allowing for a quicker deployment and saving some space (which is sometimes valuable/necessary to do so, depending upon where the application is going to be deployed).

craw-daddy/wells-202306

Important notes

Remark