Add a reducer server for performing aggregation operations.

[gibber9809]

Description

Adds a heavily refactored version of the reducer server from the prototype distributed setup as well as the code necessary to construct local in memory reduction pipelines.

Validation performed

• Tested that a version of the reducer server from earlier in this PR's lifetime works correctly with the prototype clp package setup.
• Tested that new networking code in this reducer server interacts correctly with a dummy python search scheduler implementation
• Validated that the current version of the reducer works correctly end-to-end after integrating with current clp package

[kirkrodrigues]

An incomplete review but something to start with.

Besides the inline comments (in order of priority):

• Add class docstrings.
• Fix most (if not all) of the clang-tidy warnings (using the clang-tidy config I gave you).
• Add method docstrings for non-trivial public methods.
• Add method docstrings for non-trivial private methods.

[kirkrodrigues]

Finished a pass on the code. Can you address the unresolved comments and then test everything?

[gibber9809]

Finished a pass on the code. Can you address the unresolved comments and then test everything?

Tested that everything works end to end after my most recent set of changes.

[kirkrodrigues]

We need to catch the exceptions here or in main, right? Otherwise main will throw which is no bueno. My preference would be to catch them here like we do in the other programs.

[kirkrodrigues]

Commit msg: Add a reducer server for performing aggregation operations. (#221)?

Dropped "distributed" since it might imply that the reducer itself is distributed, but no big deal if you want to keep it in.