feat: make dedupe faster at I/O level

[alpha-ulrich]

Summary:
Uses Arc to ensure that dedupe requests do not generate new clones of the response, but rather point to the same response using the Atomically referenced counter.

Issue Reference(s):
Fixes #3028

/claim #3028

Build & Testing:

• I ran cargo test successfully.
• I have run ./lint.sh --mode=fix to fix all linting issues raised by ./lint.sh --mode=check.

Checklist:

• I have added relevant unit & integration tests. [Modified relevant tests after code changes]
• I have updated the documentation accordingly. [No update required for documentation]
• I have performed a self-review of my code.
• PR follows the naming convention of <type>(<optional scope>): <title>

[codecov]

Codecov Report

Attention: Patch coverage is 88.23529% with 6 lines in your changes missing coverage. Please review.

Project coverage is 86.32%. Comparing base (e3a6026) to head (d3850c3).

Files with missing lines	Patch %	Lines
src/core/data_loader/dedupe.rs	85.18%	4 Missing ⚠️
src/core/ir/eval.rs	75.00%	1 Missing ⚠️
src/core/ir/eval_io.rs	92.30%	1 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #3237      +/-   ##
==========================================
- Coverage   86.32%   86.32%   -0.01%     
==========================================
  Files         282      282              
  Lines       28798    28821      +23     
==========================================
+ Hits        24860    24879      +19     
- Misses       3938     3942       +4     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[tusharmath]

@alpha-ulrich Thanks!
Can you add some e2e benchmarks to see if there is an observed difference in performance? We eventually end up cloning the data in the PR, so I am curious to see if Arc is making a difference in performance.

[alpha-ulrich]

@tusharmath sure, I'll see what can be done, later today.
I'll have a look at the graphql-benchmarks repository, to see if I can reuse some parts from there to accomplish this.

If there are any specific metrics that you would want reported, do let me know.