-
Notifications
You must be signed in to change notification settings - Fork 61
Developer guide
Interested in contributing to 4CAT or extending it with your own modules? The following pages offer the information you need to do this:
- Architecture: A high-level overview of 4CAT's software architecture.
- How to make a data source: On adding support for data capture from new platforms.
- How to make a processor: On adding new analytical processors to 4CAT.
- Input fields: On the API to use for connecting data sources and processors to the web interface.
- Using git to track changes, create a new branch or fork is the best way to developer 4CAT. The Docker 4CAT containers inherit the git branch you cloned and so you can use all of git's functionality inside the containers.
- Logs are stored by default in the
logs
folder (/usr/src/app/logs
inside Docker). In Docker, this folder is a shared volume so it is accessible/identical via both4cat_backend
and4cat_frontend
(as aredata
andconfig
). - Commands to control 4CAT's backend:
python3 4cat-daemon.py status
python3 4cat-daemon.py stop
python3 4cat-daemon.py start
Docker's containerized approach allows you to easily change 4CAT and its environment and redeploy your local application. You can make changes locally in your cloned 4CAT repository, shutdown your running 4CAT instance with docker compose down
, and then recreate the application with docker compose -f docker-compose_build.yml up --build -d
. This will rebuild your application with your local files following the docker-compose_build.yml
file (ensure you use that file as opposed to docker-compose.yml
since this file pulls our pre-built images directly from Docker Hub).
Note: if you wish to remove all collected data as well, use docker-compose down -v
which removes the volumes as well.
Often it is not necessary to rebuild the entire application. If you only change certain files or create new processors/datasources, you can copy those files directly into your Docker 4CAT containers. docker cp path/to/file 4cat_backend:/usr/src/app/path/to/file
will copy a file to the 4cat_backend
container. Remember to also copy the same files to the 4cat_frontend
container which works as a mirror to the backend. The frontend will generally pick up all changed files automatically so long as you refresh your browser, however, the backend will need to be restarted. You could restart the container in Docker desktop or run docker exec 4cat_backend python3 4cat-daemon.py restart
.
You can also connect directly to the containers with either the Docker GUI or the command docker exec -it 4cat_backend /bin/bash
where you can run commands, test your python3 environment and files, and even edit files directly. (You could install your preferred text editor if you wish with a command such as apt-get install nano
directly in the container.)
- 4CAT Docker containers do not need to be running to copy files to and from them (helpful if an error causes a container to crash!)
-
docker compose
essentially relies on thedocker-compose.yml
(ordocker-compose_build.yml
) file and your.env
file, but also uses thedocker\Dockerfile
to build the 4cat image environments and thedocker/docker-entrypoint.sh
file which runs every time the4cat_backend
container starts. - You can open additional ports by adding them to
docker-compose_build.yml
. For example, if you wished to directly connect to the database from your local machine, just add the lineports: - 5432:5432
to thedb
container underservices
and recreate 4CAT allowing you to connect to the database from your local machine.
- View container logs
docker container logs container_name
- Stop running container
docker stop container_name
- Start stopped container
docker start container_name
- Remove container
docker container rm container_name
Useful to remove then recreate with new parameters (e.g. port mappings) - Remove image
docker image rm image_name:image_tag
Useful if you need to changeDockerfile
ordocker-entrypoint.sh
and rebuild- Note: must also remove any containers dependent on image; you could alternately create a new image with a different name:tag
- Copy files into container
docker cp path/to/file container_name:/full/path/to/container/directory/
Can update and change files (e.g.config.py
or other configuration files) Note: may require restarting the container to take effect
Documentation, in this documentation branch, is done (semi-)automatically, with the use of docstrings formatted in RestructuredText, through Sphinx (v4.5) and its autodoc and autosummary modules. Therefore, when merging commits and general updates to the code into this branch, one ought to be aware that the overall code structure have changed slightly; mostly in terms of directories and filenames being renamed with underscores taking the place of hyphens throughout. This is as Sphinx need to import all relevant directories and py-files as modules and packages. This is also why there are a lot of empty "init.py" files.
Currently many functions lack thorough documentation, or any documentation at all, please follow the steps below to update as you encounter unfinished documentation!
Docstring format:
Please follow the official ReStructuredText formatting for your docstrings:
"""
[Summary]
[Detailed description if needed]
:param [ParamName]: [ParamDescription], defaults to [DefaultParamVal]
:type [ParamName]: [ParamType](, optional)
:raises [ErrorType]: [ErrorDescription]
:return: [ReturnDescription]
:rtype: [ReturnType]
"""
Make sure to, as a minimum, keep the empty lines between summary and param-list
How to update the documentation
For more info, this short guide might be helpful.
Make sure that you have sphinx
, sphinx-mdinclude
sphinx-rtd-theme
installed via pip in your environment:
$ (sudo) pip install sphinx
$ (sudo) pip install sphinx-rtd-theme
$ (sudo) pip install sphinx-mdinclude
Scenario: I changed or added new information to existing docstrings:
- Update project files and merge commits into this documentation branch, be aware of possible conflicts within naming conventions and resolve accordingly
- Open the "docs" folder in your terminal
- Run
make clean && make html
as this will clear existing HTML and regenerate new, including your new docstrings, and you can do what you want with the newly generated HTML-files in the /build directory - success!
Scenario: I added a new submodule (i.e. a new datasource)
-
Update project files and merge commits into this documentation branch, be aware of possible conflicts within naming conventions and resolve accordingly
-
Open the "docs" folder in your terminal
-
Either:
- Delete the rst file(s) in question within the "source"-directory (leave the index.rst!)
or
- Manually add the stub in the relevant rst file within the "source"-directory as required
-
Regenerate missing rst files by running
sphinx-apidoc -o ./source .. "/*setup*" "/*4cat-daemon*" "/*config*"
in your terminal and reformat accordingly for layout, as all formatting will be lost upon deletion. The last part of the command lists the aspects of the code that we want to exclude, as these are work poorly for Sphinx documentation. -
run
make clean && make html
as this will clear existing HTML and regenerate new, including your new docstrings, and you can do what you want with the newly generated HTML-files in the /build directory - success!
Documentation Troubleshooting
If you encounter errors, make sure your current python environment (venv?) has the packages installed from pip which 4cat is listing in setup.py. If you have issues installing psycopg2==2.8.6
then try to
pip install psycopg2-binary==2.8.6
instead.
Flask needs to be executable from Sphinx' end, so it might ask to have FLASK_KEY filled. In that case make a config.py
from config.py-example
and populate the secret name (see this on how to generate one)
🐈🐈🐈🐈