Pass tokens via env variable or config file

[sbliven]

Summary

To avoid putting tokens or credientials in scripts and jupyter notebooks, I usually parse them from a config file. It would be nice to move this code inside pyscicat.

Current behavior

Creating a ScicatClient currently requires either a token or a username/password to be supplied. Providing only the base_url results in an AssertionError.

Proposed behavior

If neither token nor credentials are supplied, read them from the following sources:

1. $SCICAT_TOKEN environmental variable
2. $SCICAT_USER and $SCICAT_PASSWORD
3. The file $XDG_CONFIG_HOME/scicat/config.ini (path subject to discussion) containing either of the above variables.

For instance, I might set /.config/scicat/config.ini to the following:

SCICAT_USER = ingestor
SCICAT_PASSWORD = myp4ssw0rd

[sbliven]

There are a few things open for discussion:

• Spelling of the variables and paths
• Format for the config file (ini, toml, bash-like, yaml, json)

[sbliven]

I often want to switch between production and development tokens. Maybe this could also be added somehow in the config (eg different ini sections for each environment, plus an easy way to switch between them).