marshmallow: evaluate fields.Str vs SanitizedUnicode

[ppanero]

Several fields, such as type, are strings that can only take as value one of an enumeration. Therefore at the moment there is no need for them to be sanitized and could go along with just being a Str(). However, if we open these to customization (e.g. introduce your own CVs) it might cause problems (maybe in other languages that are not English).

Run some tests on performance from marshmallow.fields.Str vs invenio_records_rest.schemas.fields.SanitizedUnicode.

1 load/dump:

Sanitized: 0.0006556510925292969 seconds
Marshmallow Str: 0.0002028942108154297 seconds

1000 load/dumps:

Sanitized:  0.18001985549926758 seconds
Marshmallow Str: 0.09990668296813965 seconds

100000 load/dumps:

Sanitized: 18.859431743621826 seconds
Marshmallow Str: 10.20322585105896 seconds

[fenekku]

Grooming
This should be tackled when we deal with #5

[github-actions]

This issue was automatically marked as stale.

[fenekku]

I think this can be closed. The bottom line:

• if we know the string will never be used to hold potential internationalized string, then use Str (e.g. enums)
• otherwise use SanitizedUnicode (e.g. free text, translated text)