Optimization: Prune manifest in snapshot overwrite operations

[gabeiglio]

Rationale for this change

Doing some performance tests for overwriting partitions, we noticed that PyIceberg took double the time it usually takes java based implementation, we noticed that _exisiting_manifests does not take advantage of manifest pruning before reading all Manifest Entries

In this PR I:

• Moved methods from _DeleteFiles to _SnapshotProducer parent class to share with other classes (_OverwriteFiles)
• Implemented manifest pruning over all deleted files partitions to not read manifests that do not match file partitions
• Refactored the method to only iterate once over all files (instead of multiple)

Are these changes tested?

I believe current tests in tests/integration/test_writes.py cover all cases

Are there any user-facing changes?

Nope

[geruh]

Thanks for raising this @gabeiglio! I went through the core behavior here and tested out some of the pruning logic. but ultimately, the existing integration tests like test_overwrite_partitioned_table, and test_delete_overwrite would catch regressions. Just a few inline comments from my side.

[geruh]

This moved a bunch of the imports to multi line for some reason

[gabeiglio]

its the linter, so every commit the pre-commit hook will flatten it again

[geruh]

yeah looks cosmetic since it's linting alr committed code

[Fokko]

This makes sense to me @gabeiglio Thanks for fixing this 👍

[geruh]

LGTM!

[Fokko]

Let's goo! Thanks @gabeiglio for working on this, this is super valuable. Thanks @geruh and @yingjianwu98 for the review 🚀