Support configuration in multiple ways

[Peter9192]

• default config
• override based on (partial) config file in standard location
• override based on (partial) config file in location passed through CLI/API
• override individual options through CLI or API

Perhaps use hydra https://hydra.cc/docs/intro/

[sverhoeven]

Now not ideal, as config fields are used on import time. For example at https://github.com/phenology/springtime/blob/ce3bdfd250b3439fe18f78914291147aba0531a2/src/springtime/datasets/phenocam.py#L23

You need to change configuration before importing things that use config.

[sverhoeven]

Use https://pypi.org/project/xdg-base-dirs/ to get cache and config dir

[sverhoeven]

springtime --output-dir /tmp/outputdir --cache ~/.cache/springtime --config ~/.config/springtime <recipe>
# for example ~/.config/springtime/pep725_credentials.txt

[sverhoeven]

I would like to have several options for output location:

1. output in specific folder
   • With --output-dir /tmp/myoutput so data.csv is written in /tmp/myoutput/data.csv`.
2. output in current working directory.
   • With --output-dir . so data.csv is written in ./data.csv.
3. output in sub directory like the current time of specific folder. for example /tmp/springtime-jobs/springtime-24072023T14:18:23/data.csv
   • can configure /tmp/springtime-jobs with --output-root-dir.

If --output-dir is absent then <output-root-dir>/springtime-24072023T14:18:23/data.csv If --output-root-dir is absent then it is ..