Revisit relationship between Distribution and PathDistribution

Should Distribution objects expose a method to resolve paths?

Yes, that is one solution. There is already .locate_file() which is not far off, except that its implementation in PathDistribution precisely removes the metadata directory that we need in this case.

More generally, I wonder if there is just a tad too much coupling between the Distribution and PathDistribution classes?

There is a mix of abstract and concrete methods in the Distribution base class, and several of the concrete methods seem fairly tightly bound to the PathDistribution subclass. For example, the at() static method creates and returns a PathDistribution instance outright, and the metadata(), entry_points(), files(), and requires() methods are all implemented with assumptions as to which files should be available and that the organization should resemble a dist-info or egg-info distribution.

Now, there is (AFAICS) only the one (PathDistribution) subclass of Distribution available in this project, and I don't really have enough experience with this project or its users to really imagine what other Distribution subclasses would or could exist in the wild.

Still, I would suspect that a subclass that was sufficiently different from PathDistribution to rather prefer subclassing Distribution directly, would then find itself not merely implementing the read_text() an locate_file() abstract methods, but would probably also need to _re_implement several of metadata(), entry_points(), files(), or requires() as well.

Hence, I would raise the question whether some of these methods would be better off with their concrete implementations moved into the PathDistribution subclass? (Of course leaving abstract methods behind in Distribution where that makes sense.) This would grant these concrete methods direct access to the stuff they need inside PathDistribution, instead of having to add more interfaces to Distribution for stuff that really only makes sense for PathDistribution.

Still, this is a much bigger refactoring than I set out to do, and I would not feel comfortable starting this without consulting you.

Originally posted by @jherland in https://github.com/python/importlib_metadata/pull/437#discussion_r1141220985

Start by reading the Distribution and PathDistribution classes and the methods named in the issue: at(), metadata(), entry_points(), files(), requires(), read_text(), and locate_file(). Review the linked pull request discussion for the original context. Done would require an agreed design for separating subclass-specific behavior and an implementation that preserves the relevant distribution functionality.