bug: Filter predicate is duplicated in `TableScanBuilder` and `TableScan` 🤦🏼‍♂️

[sdd]

Somehow, between my PR and @viirya's PR getting merged, we've ended up with the filter predicate being duplicated in the TableScanBuilder and the TableScan.

See predicates and filter:

iceberg-rust/crates/iceberg/src/scan.rs

Lines 52 to 56 in adc5d20

	predicates: Option<Predicate>,
	snapshot_id: Option<i64>,
	batch_size: Option<usize>,
	case_sensitive: bool,
	filter: Option<Predicate>,

iceberg-rust/crates/iceberg/src/scan.rs

Lines 85 to 91 in adc5d20

	/// Specifies a predicate to use as a filter
	pub fn with_filter(mut self, predicate: Predicate) -> Self {
	// calls rewrite_not to remove Not nodes, which must be absent
	// when applying the manifest evaluator
	self.filter = Some(predicate.rewrite_not());
	self
	}

iceberg-rust/crates/iceberg/src/scan.rs

Lines 99 to 103 in adc5d20

	/// Add a predicate to the scan. The scan will only return rows that match the predicate.
	pub fn filter(mut self, predicate: Predicate) -> Self {
	self.predicates = Some(predicate);
	self
	}

And here in TableScan, bound_predicates and filter:

iceberg-rust/crates/iceberg/src/scan.rs

Lines 192 to 196 in adc5d20

	bound_predicates: Option<BoundPredicate>,
	schema: SchemaRef,
	batch_size: Option<usize>,
	case_sensitive: bool,
	filter: Option<Arc<Predicate>>,

The predicate pushdown uses one of these but the file plan uses another, which caused a lot of confusion when I was trying to track down why my queries were misbehaving!

[sdd]

@liurenjie1024 can you add the good first issue label please?

[liurenjie1024]

@liurenjie1024 can you add the good first issue label please?

Sure, done.