[Datasets] [Hotfix] Add ray_remote_args to read_text.

[clarkzinzow]

Adds ray_remote_args to ray.data.read_text(). This exists on all other ray.data.read_* APIs, so it appears that read_text() was missed.

Checks

• 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 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 :(

[stale]

This pull request has been automatically marked as stale because it has not had recent activity. It will be closed in 14 days if no further activity occurs. Thank you for your contributions.

• If you'd like to keep this open, just leave any comment, and the stale label will be removed.

[clarkzinzow]

Will be coming back to this in the next week or so.

[krfricke]

I've taken a look.

• The test is stochastic, as sometimes the remote tasks are scheduled only on the bar node (as expected)
• There were two sources of errors
  1. We didn't pass the remote args to flat_map(), which were sometimes scheduled on the foo node
  2. ds.take() also schedules remote tasks, which were sometimes scheduled on the foo node
• To resolve this, we're now passing remote args to flat_map()
• I don't think we need to solve ds.take() - but we need to move that assertion to the bottom of the test.

This all passes locally for me every time now.

See these changes: 7b4ee95

[krfricke]

Bit of a self-accept but still - cc @clarkzinzow to merge

[clarkzinzow]

@krfricke Thanks a million for finding the blocker bug! Definitely a 🤦 moment for me. 😁

I think this is good to merge, with a P2 follow-up around adding a TextDatasource to get rid of this two-stage read.