An IndieWeb original micropub server, written in Python Flask.
chaos-micropub began as an IndieWeb micropub server and retains the name, but to get started it's locked into a specific Hugo workflow. To me the capacity to publish personal comments to my site is more important than interoperability. It does not presently do anything particularly IndieWeb-like, except perhaps by abiding by the rule to eat one's own dogfood.
This is the interface for creating a note. There are more fields as you scroll down, but I like ample whitespace.
This is the interface for editing a log.
The entire workflow can be accomplished via Make, however, you'll need to configure the scripts on which Make depends by creating a .env file. env-example supplies the required variables.
This code flows through three steps:
Develop -> Verify -> Deploy
In develop, the goal is to provide quick feedback for code changes and test this service as a single system.
In verify (or UAT), the goal is to confirm that this service is operable with its other services. It mimics the production environment.
In deploy, the goal is to deliver this service to my production server.
The processor architecture and container tooling I use in production differs from what I use locally. To avoid issues porting this service to the production server, I use the same server to conduct my UAT and production builds and tests.
Docker image: acbilson/micropub-dev:alpine Hugo config: config-dev.toml Hugo build script: build-site.sh Hugo theme: acbilson/chaos-theme Markdown content: acbilson/chaos-content
- Docker
- Make
- bash
- envsubst
- entr (optional)
To build a local Docker developer image, run:
make build
The developer image runs the Flask application directly without UWSGI to take advantage of file change refresh. Also, authentication is disabled in the developer image so I don't need a fresh OAuth token for local development.
To start the developer container, run:
make start
Unit tests are run on the developer image with the following command:
make test
When I'm making changes, I like to get immediate feedback from my unit tests on save. This can be accomplished with the invaluable entr command. It will watch for file changes and run whatever command you'd like after any change happens. Here's how I use it:
find ./src | entr make test
Docker image: acbilson/micropub-uat:alpine Hugo config: config.toml Hugo build script: build-site.sh UWSGI config: micropub.ini Hugo theme: acbilson/chaos-theme Markdown content: acbilson/chaos-content
A production-like uat server with:
- Podman
- OpenSSH
- bash
To build a Docker verify image on the UAT server, run:
make build-uat
To deploy the image on the UAT server, run:
make deploy-uat
The UAT image runs the Flask application behind UWSGI to mimic the production instance.
Docker image: acbilson/micropub:alpine Hugo config: config-uat.toml Hugo build script: build-site.sh UWSGI config: micropub.ini Systemd service file: container-micropub.service Hugo theme: acbilson/chaos-theme Markdown content: acbilson/chaos-content
A production server with:
- Podman
- OpenSSH
- bash
- systemd
- A web proxy like Nginx to broadcast this service to the public Web
To build a Docker production image on the prod server, run:
make build-prod
To start the production service, run:
make deploy
The production image runs the Flask application behind UWSGI, which allows for multi-threading and interfaces with my web proxy. The container is managed by systemd.
Since my app is not production-critical I will sometimes publish hotfixes to production before testing in uat. The deployment process is greatly simplified with redeploy, which disables, rebuilds, and enables my service.
make redeploy