[Data] Deprecate Reader

[bveeramani]

Why are these changes needed?

tl;dr Reader adds unnecessary complexity, so we're removing it.

We're deprecating Reader and moving its methods into Datasource. There are several reasons why:

1. Reader introduces a pass-through layer.

class FileBasedDatasource(Datasource):
    def create_reader(self, **kwargs):
        return _FileBasedDatasourceReader(self, **kwargs)

2. Reader subclasses contain circular dependencies.

3. Datasource parameters are specified with implicit (and error-prone) keywords arguments. For example:
   

   ray/python/ray/data/datasource/csv_datasource.py

   Lines 35 to 38 in 8504563

   	read_options = reader_args.pop(
   	"read_options", csv.ReadOptions(use_threads=False)
   	)
   	parse_options = reader_args.pop("parse_options", csv.ParseOptions())

Going forward, you would implement Datasource.get_read_tasks and Datasource.estimate_inmemory_size directly rather than Reader.get_read_tasks and Reader.estimate_inmemory_size.

This change is part of a larger effort to clean up the Datasource-related interfaces. For more information, see the design document.

Related issue number

Towards #40296

Checks

• I've signed off every commit(by using the -s flag, i.e., git commit -s) in this PR.
• I've run scripts/format.sh to lint the changes in this PR.
• I've included any doc changes needed for https://docs.ray.io/en/master/.
  • I've added any new APIs to the API Reference. For example, if I added a
    method in Tune, I've added it in doc/source/tune/api/ under the
    corresponding .rst file.
• I've made sure the tests are passing. Note that there might be a few flaky tests, see the recent failures at https://flakey-tests.ray.io/
• Testing Strategy
  • Unit tests
  • Release tests
  • This PR is not tested :(

[stephanie-wang]

Can you provide some context in the PR description?

[stephanie-wang]

Looks good, modulo the TODOs!