feat: Implement abstract EvaluationHarness class

[shadeMe]

This is a part of a series of PRs that revolves around implementing an evaluation harness (c.f deepset-ai/haystack#7526).

EvaluationHarness provides the base interface for class that implement usecase-specific evaluation harnesses. It's meant to be generic over its inputs, overrides and outputs. A key feature of this abstraction is the support for overriding the initial parameters of arbitrary components in the evaluated (and evaluation) pipeline. It leverages the Pipeline abstraction's serialization hooks to transparently instantiate new pipeline internally for each set of overrides.

No unit tests are provided as this class does not implement any new functionality.

Blocked by #2.

[coveralls]

Pull Request Test Coverage Report for Build 9189156861

Details

• 20 of 38 (52.63%) changed or added relevant lines in 2 files are covered.
• No unchanged relevant lines lost coverage.
• Overall coverage decreased (-6.0%) to 93.624%

---

Changes Missing Coverage	Covered Lines	Changed/Added Lines	%
haystack_experimental/evaluation/eval_harness.py	18	36	50.0%

Totals
Change from base Build 9189123832:	-6.0%
Covered Lines:	279
Relevant Lines:	298

---

💛 - Coveralls