GH-37657: [JS] Run bin scripts with ts-node

[trxcllnt]

• update bin scripts to use ts-node in transpile-only mode
• remove ts-node config from top level tsconfig.json
• fix tsc errors in bin scripts

Fixes #37657

• Closes: [JS] [Benchmarking] JavaScript microbenchmarks are failing #37657

[github-actions]

⚠️ GitHub issue #37657 has been automatically assigned in GitHub to PR creator.

[domoritz]

Looks good overall. Eventually, we could even try running the benchmarks in e.g. bun so that we also have jscore as a comparison and not just v8 via node.

[kou]

It seems that integration CI job failure is related to this: https://github.com/apache/arrow/actions/runs/6152750564/job/16695430388?pr=37668

Could you check the log?

[austin3dickey]

I just noticed that this open PR is expected to close #37657. That's great news! Could someone please take another look to see if we can merge and turn the benchmarking jobs green again?

[kou]

@trxcllnt Could you rebase on main and resolve conflict?

@domoritz Could you review this again after that?

[domoritz]

I'm happy to do another review.

[raulcd]

Hi, from my understanding benchmarks haven't been run for the last month or so for JS. It would be nice to re-enable them to validate we are not adding any performance issue on the Release.

[austin3dickey]

@ursabot please benchmark lang=JavaScript

[ursabot]

Benchmark runs are scheduled for commit 2ee36ea. Watch https://buildkite.com/apache-arrow and https://conbench.ursa.dev for updates. A comment will be posted here when the runs are complete.

[conbench-apache-arrow]

Thanks for your patience. Conbench analyzed the 0 benchmarking runs that have been run so far on PR commit 2ee36ea.

None of the specified runs were found on the Conbench server.

The full Conbench report has more details.

[austin3dickey]

I checked out this branch locally, rebased to main, and the JavaScript benchmarks work again. I don't think I have access to push to this fork, but that's a great sign! Thanks for the work so far.

If we don't merge this PR soon, I'm going to turn off the automatic benchmarking for JS since it's failing every time and covering up other errors like voltrondata-labs/arrow-benchmarks-ci#166.

[assignUser]

I have rebased the PR locally favoring main for the yarn.lock conflicts. In the interest of getting the benchmarks going again I have opened an alternative PR #38500 (keeping authorship for @trxcllnt of course).

[raulcd]

Should we close this PR as the other one has been merged?

[assignUser]

Yep, sorry I forgot to do that.