Running locally

Python 3.11 or higher is required to run. Dependencies and virtual environment is managed using Poetry

Install dependencies into virtual environment

poetry install

Create a .env file by copying from .env.template:

cat .env.template > .env

Change any parameters in .env to configure your specific run. Note: the .env file is intended for local setup and might include secrets. It should not be checked into git. Also, make sure that no secrets are added to the .env.template file.

Run

poetry run python .\fos_forecast.py

Optionally, specify the time as a command line argument

poetry run python .\fos_forecast.py --timestamp '2023-12-15T00:00:00+01'

Testing requirements file for cloud environment

To test whether the script is runnable in the cloud environment, try to run the script in an azure functions docker container.

Build image

docker build -t tmpslopes .

Run image in interactive mode

docker run -it tmpslopes bash

Inside image, run script

python fos_forecast.py

Note that this container does not have access to blob storage or the P drive. It uses local temporary input files downloaded from blob storage in a previous local run.

Deploying code to production

Check access to azure devops

Check the pipelines overview in azure devops project

"Run pipeline", and select the "Publish app (production environment)" checkbox.

Manually triggering azure timed function

Get the function url and _master key from the Azure portal "App Keys" pane for the azure function, then run:

curl -kv -XPOST "<function_url>/admin/functions/TimedTrigger" -H "x-functions-key: <master_key>" -H "Content-Type: application/json" -d '{"input": ""}'

A 202 response denotes a successful request. .