Include author/maintainer email information in the schema

[Lance-Drane]

We can inspect the project metadata (defined in pyproject.toml under [project.authors] and [project.maintainers] but defined elsewhere once packaged) with importlib and include it in the schema. Note that this information will have to be optional.

Things to consider:

• we may need to include a flag in the configuration to allow for users to opt-out of including their information in this schema, this would be an "all or none" approach.
• we can scan this information on either the Service itself, or do this per capability; we can also do both:
  • doing this on the Service is trivial and could be done at schema generation
  • doing this on the Capability would require we be able to inspect the package the Capability originated from

[Lance-Drane]

To get this information:

SAMPLE_PACKAGE = 'pydantic'

from importlib.metadata import metadata
metadata(SAMPLE_PACKAGE).get('Author-email')

If information was provided, this gives back a single string to parse, with emails separated by commas:

' Lance Drane <Lance-Drane@users.no-reply.github.com>, Someone Else <another-email@somewhere.com>'

Note that this can give back garbage if the characters aren't ASCII, i.e. fastapi returns '=?utf-8?q?Sebasti=C3=A1n_Ram=C3=ADrez?= <EMAIL_REDACTED>' instead of Sebastián Ramírez. To fix this, this sort of works:

from email.header import decode_header
from importlib.metadata import metadata

authors = metadata('fastapi').get('Author-email')
if authors:
   print(decode_header(authors))

#### OUT ####
[(b'Sebasti\xc3\xa1n Ram\xc3\xadrez', 'utf-8'), (b' <EMAIL_REDACTED>', None)]

The second value of each tuple in the list tells how to decode the bytes (or None, if not needed), so for example 'Sebasti\xc3\xa1n Ram\xc3\xadrez' will correctly display as Sebastián Ramírez in UTF-8 format