sudo apt-get install docker.io
sudo adduser $(whoami) docker
You need to sign out and sign back in again for changes to take effect.
Makeradmin uses Python 3.11.
sudo apt-get install python3.11-dev python3.11-doc python3-pip
sudo apt-get install npm
Clone this git repository to a suitable place on your computer / server
Tip
Start by initializing and activating a virtual python environment in the cloned folder. This makes sure that the packages used for Makeradmin are isolated into its own "environment" (i.e. in the local directory), and will not interfere with packages already installed.
python3 -m venv .venv
source .venv/bin/activate
Run the firstrun script
make firstrun
This will build docker images and configure the database. This may take quite some time.
It will also generate a .env
file with new random keys and passwords that the system will use.
You will be prompted for if you want to create a new admin user, it is recommended to do this here.
You will also be prompted for if you want to create some fake members, transactions etc which can be useful for development.
If you are deploying on a server you need to configure hosts and other settings by editing the .env
file.
If you do modify the .env
file you need to restart the services afterwards by running
docker-compose up -d --build
Run all services locally (but you will have to insert data, see below):
make run
You can also run in dev mode where source directories are mounted inside the containers and sources are reloaded when changed (in most cases):
make dev
This can be done from the web UI, but it can be convenient to do it from the commandline too
python3 create_user.py --first-name "Maker" --last-name "Makersson" --email "maker@example.com" --type admin
Before running the command above you might have to run the commande below to install all dependencies.
make init
If the admins don't seem to have the permissions that they should have (possibly because you have upgraded makeradmin to a newer version) then you might have to update the permissions. Simply running the firstrun script again will do this:
make firstrun
Go to:
Go to the member page and fill in the email address corresponding to the user created previously. A link will then be printed in the terminal (where make dev
is run) that allows you to login. E.g.
[...]
public_1 | 10.0.2.2 - - [18/Dec/2018:20:50:23 +0000] "GET / HTTP/1.1" 302 223 "http://localhost:8011/member/login/XHCgGQZGrjuG6bO7TVPkikTfQVRo6Eqn?redirect=%2Fmember" "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/70.0.3538.110 Safari/537.36"
[...]
Some pages need special products in order to function (e.g. the registration page, or the member page). The products can be created using make firstrun
.
The .env
file includes a number of variables that are unset by default.
If you want emails to be sent, you'll need to set the MAILGUN_DOMAIN
, MAILGUN_KEY
and MAILGUN_FROM
variables.
You will also want to set the ADMIN_EMAIL
variable to some mailbox that you monitor.
To deploy on any host which is not localhost, you'll need to change the HOST_BACKEND
, HOST_FRONTEND
and HOST_PUBLIC
variables.
These are important to make sure links work, but also to handle CORS in the browser.
Systests are written in python and the sources for the systests are in the api/src/systest directory (because it shares a lot of code with the api unittests). There are tests using the api as well as selenium tests. Those tests are also run in Github actions.
You can run the tests in test containers using a one off db with:
make test
To run a single test, you can pass options to pytest using the PYTEST_ADDOPTS
environment variable, for example:
PYTEST_ADDOPTS='-k "test_empty_cart_fails_purchase"' make test
Other test options that pytest supports can be passed using the same variable.
Or you can run against your local running environment with:
make dev-test
And you can also run single tests against your local running environment using you favorite test runner (like pytest).
The api directory also contains unittests that can be run standalone, they will also run when running make test
.
Javascript unit tests are run when the images is build but they can also be run against the source directly
by make test-admin-js
or npm --prefix admin run test
.
If you for some reason want to remove the existing database and start from scratch you can run the command
make clean-nuke
Warning
This will completely wipe out all your makeradmin data!
After this you can run make firstrun
again to set things up again.
Create your own stripe account and add your keys to the .env
file to allow purchases to work.
You will not be able to go to the checkout unless you have a Stripe key in the .env-file. If this is set up, you can use Stripe's fake cards for completing the purchase.
Makeradmin is used as the truth, and Stripe is automatically synced with the makeradmin products.
To handle subscriptions properly, the server needs to listen to stripe webhooks and configure subscription products (see next session). You can do this by installing the Stripe CLI and forward events using
stripe listen --forward-to http://localhost:8010/webshop/stripe_callback
After the forwarding has started, you'll need to copy the signing secret it gives you, and put it in your own .env
file in the key STRIPE_SIGNING_SECRET
.
Note: When running tests, this is not necessary, as it will poll the stripe server automatically.
For accounting Makeradmin supports exporting the transaction history as sie files.
The transaction history is compared with Stripe to make sure it matches and there are no extra transactions in Makeradmin or Stripe.
Debit bookings adjust the product price with the transaction fee. However, each transaction usually contain multiple products. Thus, the transaction fee neds to be distributed over the products if the debit accounts or cost centers for the product are different. This adjustment involves distributing the transaction fee among the products in proportion to their total price within the transaction, which takes into consideration the quantity of products purchased. Any rounding errors are corrected by adjusting them to the nearest, higher or lower, amount. Any left over cent is added or removed on the most expensive price to make sure the total sum is correct.