This repository holds the code for wikispeedruns.com.
- Python 3.7 or greater
- MySQL Server 8 Download here
We recommend creating a Python virtual environment for running the server.
python -m venv env
For Windows Powershell:
./env/Scripts/Activate.ps1
For Linux
source env/bin/activate
Then install the requirements (with your virtual environment activated)
pip install -r requirements.txt
There are a number of scripts to help setup the web app in scripts.
Once the MySQL server is running, you will need to create an account. By
default we assume an account user with no password (see
default.json). If you wish to use a different MySQL
setup, you can create prod.json with the relevant MySQL fields in
config which will override default.json.
Then create the database and tables using the provided script.
cd scripts
python create_db.py
There is also an interactive script (with instructions in the scripts) which
can be used to set up a local admin account. Through the admin account,
prompts can be managed through /manage.
cd scripts
python create_admin_account.py
(Optional) Finally, there is also a script to populate the database with data for local development
cd scripts
python populate_db.py
Set the environment variable FLASK_ENV in whaterver command prompt you use plan to use
for running the flask server. This will allow the local instance to reload automatically
when files are changed.
For example, in Linux
export FLASK_ENV="development"
Or in Windows Powershell
$env:FLASK_ENV="development"
From the top level directory
flask run
In order to run the tests locally, you need to create another account in MYSQL
with username testuser and password testpassword. Our tests are configured to run
against this account by default.
Then, simply run pytest from the test directory.
cd test
pytest
Note that these tests are also run in Docker upon making a PR using Github workflows. In the future we may setup docker to run tests as well.
The asynchronous task queue for scraper tasks are supported by 2 extra tools, celery
and redis. Celery is installed as a python requirement, but redis (https://redis.io/)
needs to be installed and run separately (similar to the SQL server, see website
for instructions). The scraper task_queue also requires the scraper_graph,
which can be downloaded locally (contact one of the maintainers)
Rather than computing the path as part of the request, which freezes up the server, flask passes off the scraper tasks to another process managed by celery (and communicates through redis). These tasks are defined using python decorators, examples of which can be seen here.
Unfortunately, neither celery nor redis are supported on windows. So if you have a windows development machine, you will have to run the server through WSL. Note that if you want to keep your windows MySQL instance, you need to figure out which port the host windows machine is exposed on in WSL. See this Stack Overflow post. Note that this changes everytime WSL is restarted.