This document has the following sections:
First, clone the repo:
git clone [email protected]:cc-ai/floods-backend.git
From the root of the repository, run the following to install the application dependencies into your current Python environment:
pip install -r requirements.txt
From the root of the repository, run the following to run all of the tests:
make test
For development, the best way to start the server is probably the following, which has hot reloading (which works almost most of the time):
make develop
For a more production-ready server, run the following, which will use gunicorn to launch n worker processes running the webserver such that n is the number of local CPU cores:
make serve
To build a container which can run the application server, use the following Make command:
make container
Under the hood, this will run a docker build command with the appropriate container tag, environment variables, etc. If the command is successful, a container name will be printed out at the end. You might see something like:
Step 19/19 : CMD gunicorn -w $WORKERS -b 0.0.0.0:80 ccai.bin.webserver:app
---> Using cache
---> 5aac2c0f7c37
Successfully built 5aac2c0f7c37
Successfully tagged gcr.io/climatechangeai/floods-backend:c81b86e68a78edf86a6714a497953441262470de
You can then take this container name (the part after "Successfully tagged") and run it as such:
docker run \
-e WORKERS=$(sysctl -n hw.ncpu) \
-p 5000:80 \
gcr.io/climatechangeai/floods-backend:c81b86e68a78edf86a6714a497953441262470de
As you can see here, you are also able to map port 80 in the container to port 5000 locally as well as specify the number of worker processes to use via the WORKERS environment variable.
First you must ask a member of the core team for the key and certificate files that are described here. Once you have them, create the Kubernetes secret with the following command:
kubectl create secret tls api-climatechangeai-org-tls \
--namespace ccai \
--key=./api.climatechangeai.org.key \
--cert=./api.climatechangeai.org.crt