[python] Add block-based manifest Avro reader

[XiaoHongbo-Hope]

Purpose

Speed up PyPaimon manifest reads by decoding Avro object-container blocks directly and applying partition and bucket pruning before materializing nested DataFileMeta rows. This aligns the Python path with the Java reader introduced by #9160.

Changes

• Read standard-order manifest entries by Avro block.
• Decode the partition and bucket prefix first and skip _FILE for rejected entries.
• Preserve the generic fastavro.reader() path when no early filter is present.
• Fall back to the generic reader for historical manifests with reordered top-level fields.
• Support legacy and current fastavro named-schema layouts.
• Avoid executor overhead for single-manifest reads.

Benchmark

In a fixed-snapshot planning benchmark on a large partitioned table, this reduced planning time by 33.5% (1.50x) while producing an identical scan plan.

Tests

• Manifest selective decoding, reordered fields, and named-type references
• Partition predicate regression tests
• Python 3.6 / fastavro 1.4.7 compatibility
• fastavro 1.11.1 and 1.12.2 named-schema compatibility
• BTree Java-to-Python E2E on Python 3.6.15
• Flake8 and git diff --check

[JingsongLi]

+1