feat: transaction-scoped dataloaders

[wschurman]

Why

An interesting idea from @ide: within a transaction, subsequent reads should be able to be cached locally as long as the local cache is kept consistent (not in the cache adapter though).

How

Keep a separate dataloader per-transaction that is invalidated synchronously.

Test Plan

Not tested yet.

[codecov[bot]]

Codecov Report

Merging #98 (6222f0f) into master (e3b2b23) will increase coverage by 0.03%. The diff coverage is 97.50%.

@@            Coverage Diff             @@
##           master      #98      +/-   ##
==========================================
+ Coverage   94.75%   94.79%   +0.03%     
==========================================
  Files          62       62              
  Lines        1564     1595      +31     
  Branches      171      191      +20     
==========================================
+ Hits         1482     1512      +30     
- Misses         80       81       +1     
  Partials        2        2              

Flag	Coverage Δ
integration	94.79% <97.50%> (+0.03%)	:arrow_up:
unittest	94.79% <97.50%> (+0.03%)	:arrow_up:

Flags with carried forward coverage won't be shown. Click here to find out more.

Impacted Files	Coverage Δ
packages/entity/src/internal/EntityDataManager.ts	98.73% <96.77%> (-1.27%)	:arrow_down:
packages/entity/src/EntityLoader.ts	88.00% <100.00%> (+0.32%)	:arrow_up:
packages/entity/src/EntityMutator.ts	98.03% <100.00%> (+0.03%)	:arrow_up:
packages/entity/src/EntityQueryContext.ts	100.00% <100.00%> (ø)

---

Continue to review full report at Codecov.

Legend - Click here to learn more Δ = absolute <relative> (impact), ø = not affected, ? = missing data Powered by Codecov. Last update e3b2b23...6222f0f. Read the comment docs.

[wschurman]

Abandoning for now. While this would be a small perf increase, ~98% of loads in the EAS use case are non-transactional so it would be minimal.