GH-15280: [C++] Refactor to reorganize dependencies as a prequel to moving acero out of libarrow

[ildipo]

Rationale for this change

In order to remove acero from libarrow we need to reorganize the code to reduce entanglement between compute/exec and the rest of the product by removing any dependency from the codebase (except dataset) into compute/exec.

What changes are included in this PR?

This PR removes some of the dependencies:

• some of the functions in compute/exec/util.[h,cc] that are widely used are moved to a new compute/util.[h,cc] file
• key_map and key_hash are widely used and thus moved from compute/exec to compute
• some light_array tests do use functionalities inside compute/exec, so the tests are split into 2 parts, one inside compute and another inside compute/exec

Are these changes tested?

The existing tests cover all these changes.
One additional test program has been added by splitting light_array_test in two parts.

Are there any user-facing changes?

no

• Closes: [C++] Move Acero out of libarrow #15280

[github-actions]

• Closes: [C++] Move Acero out of libarrow #15280

[github-actions]

⚠️ GitHub issue #15280 has been automatically assigned in GitHub to PR creator.

[westonpace]

This looks pretty reasonable to me. I have one suggestion regarding the test. Let me know what you want to do there and we can move forward. Thanks for helping to organize this.

[westonpace]

Hmm, it's a bit odd to have these tests split this way. Do you want to just move this entire test file to exec? Maybe later we can, in a follow-up, attempt to pull out the utilities that are needed in exec/test_util.h so the entire test file can move out of exec.

However, for now, I think it's probably ok that this is tested in exec.

[ildipo]

In the coming CR I'm moving a few more things around so I'll change this to a single file as well.

[icexelloss]

LGTM.

[icexelloss]

@ildipo Is on vacation I will double check with him and I will merge

[ursabot]

Benchmark runs are scheduled for baseline = 0582122 and contender = 44edc27. 44edc27 is a master commit associated with this PR. Results will be available as each benchmark for each run completes.
Conbench compare runs links:
[Finished ⬇️0.0% ⬆️0.0%] ec2-t3-xlarge-us-east-2
[Failed ⬇️0.45% ⬆️0.03%] test-mac-arm
[Finished ⬇️0.0% ⬆️0.0%] ursa-i9-9960x
[Finished ⬇️0.94% ⬆️0.03%] ursa-thinkcentre-m75q
Buildkite builds:
[Finished] 44edc27e ec2-t3-xlarge-us-east-2
[Failed] 44edc27e test-mac-arm
[Finished] 44edc27e ursa-i9-9960x
[Finished] 44edc27e ursa-thinkcentre-m75q
[Finished] 0582122b ec2-t3-xlarge-us-east-2
[Finished] 0582122b test-mac-arm
[Finished] 0582122b ursa-i9-9960x
[Finished] 0582122b ursa-thinkcentre-m75q
Supported benchmarks:
ec2-t3-xlarge-us-east-2: Supported benchmark langs: Python, R. Runs only benchmarks with cloud = True
test-mac-arm: Supported benchmark langs: C++, Python, R
ursa-i9-9960x: Supported benchmark langs: Python, R, JavaScript
ursa-thinkcentre-m75q: Supported benchmark langs: C++, Java

[ursabot]

['Python', 'R'] benchmarks have high level of regressions.
test-mac-arm