Add identity events for Git commits

[jjmerchante]

This PR introduces functionality to extract and eventize identities from Git commit data. It handles authors, committers, and signers, creating events for each identity found.

[sduenas]

Can we find a better regex? This one looks not helpful because we aren't capturing the email with the expression.

Also, I think we need some good examples about how people use these fields. I remember we added peer programming support on ELK long ago. Maybe you can find where we introduced it and look for the examples that we used for that.

[jjmerchante]

I used the regular expression from ELK: https://github.com/chaoss/grimoirelab-elk/blob/main/grimoire_elk/enriched/git.py#L81.

It looks like there is a single email in the format name one and name two <foo@bar>. One example is this commit: mezuro/kalibro@a9f25be.

In ELK the email is ignored. I don’t know who I should assign the email to. To both of them?

The second regular expression from ELK is incorrect; it obtains Co-author from Author and Commit fields, but Co-author is in the Commit message, not in the Author field.

[sduenas]

The original project we used to incorporate pair programming was cloudfoundy. Maybe you can have some more examples there to add to the tests.

I suggest that the email address should be part of the regular expression, maybe as an optional field. If you do this, you won't have to look for the email address later.

[jjmerchante]

I can see more examples in this archived repository: https://github.com/cloudfoundry/cflinuxfs2. They follow the same pattern: Name Surname, Name Surname and Name Surname <oneemail@domain.com>. The email sometimes comes from the first person and other times from the last.

I have improved the regular expression. The email is detected using <> and it is added as a new identity (without a name and username).

I included a test with that pattern.

[sduenas]

The original project we used to incorporate pair programming was cloudfoundy. Maybe you can have some more examples there to add to the tests.

I suggest that the email address should be part of the regular expression, maybe as an optional field. If you do this, you won't have to look for the email address later.

[sduenas]

The PR looks good. I'm suggesting minor changes, mainly in the documentation.

[sduenas]

Can you align the text and format of the table so it can be read when not rendered with a markdown reader?

[sduenas]

Suggested change

	AUTHOR_P2P_REGEX = re.compile(
	r"(?P<first_authors>.+?)\s+(?:[aA][nN][dD]|&|\+)\s+(?P<last_author>.+?)\s+<(?P<email>[^>]+)>"
	)
	# Pair programming regex. Some matching examples are:
	# - John Smith, John Doe and Jane Rae <pairprogramming@example.com>
	# - John Smith, John Doe & Jane Rae <pairprogramming@example>
	# - John Smith and John Doe <pairpogramming@example>
	GIT_AUTHORS_REGEX = re.compile(
	r"(?P<first_authors>.+?)\s+(?:[aA][nN][dD]|&|\+)\s+(?P<last_author>.+?)\s+<(?P<email>[^>]+)>"
	)

[sduenas]

LGTM