[core] Improve fixed-bucket pruning and expose scan statistics

[JingsongLi]

Purpose

PyPaimon already supports scan explanations and fixed-bucket pruning. Complete Native bucket selection using Java semantics and expose counters from ordinary planning through Plan.

Companion PyPaimon integration: apache/paimon#10609 (draft until this API reaches Rust main).

Changes

• Complete fixed-bucket pruning using Java's BucketSelector: same-field OR clauses, intersections of repeated Equal/IN constraints, per-partition predicate specialization, and the 1000-value combination bound. Keep pending and dynamic bucket assignments conservative.
• Preserve early manifest/sidecar filtering. Disable unsafe raw partition specialization under masking while retaining independent safe key conjuncts. Cache key-only selectors across partitions and use ordered lookup for large intersections; unordered literals fail open.
• Unify counters as ScanStats, requested through with_stats().plan() and retrieved through Plan.stats(). Remove tuple-returning plan_with_trace/plan_with_stats APIs. DataFusion uses the same API and statistics terminology.
• Collect lightweight statistics only. Remove ScanStatsMode, the complete flag, full-funnel fields, extra schema-statistics reads and deferred LIMIT handling. Counters describe actual planning work while retaining manifest, bucket, row-range and sidecar pruning plus LIMIT stopping.
• Keep snapshot metadata only on Plan.snapshot_id(). Preserve statistics and snapshot metadata through score attachment and audit split regrouping. Audit counters describe the final regrouped plan.
• Expose the core Plan and read-only ScanStats through thin Python bindings, including combined Delta/Changelog incremental scans. Statistics collection does not open data files.

No new dependencies or public bucket-selector API. No formats or read capabilities beyond PyPaimon's existing support.

Verification

• cargo test --locked -p paimon --lib table::: 2104 passed, 3 ignored, including 135 table-scan cases.
• DataFusion scan statistics integration: 6 passed (unrelated NaN/signed-zero case excluded).
• Python binding tests/test_read.py: 152 passed.
• Expanded PyPaimon Native regression: 501 passed, 13 skipped and 122 subtests passed; two remaining full-statistics expectations were adapted, then all 20 focused Native explain/integration cases passed.
• cargo +1.98.0 fmt --all -- --check and CI's cargo +1.98.0 clippy --locked --all-targets --workspace --features fulltext,vortex -- -D warnings passed.

Coverage includes ordinary/statistics plan equality; zero/positive LIMIT; sparse manifest/Avro sidecar decoding; pending buckets; masking and authorization; correlated partition predicates; historical/empty snapshots; cross-schema ADD/DELETE netting; Data Evolution projection; and combined Delta/Changelog windows.