perf(cbf): filter matched block txs by script relevance

Only include transactions whose outputs pay to watched scripts or whose
inputs spend known outpoints, instead of pushing every tx from a matched
block into BDK's TxGraph. On mainnet blocks with thousands of txs this
avoids significant memory and CPU overhead during wide filter scans.

Author: febyeji

src/chain/cbf.rs

@@ -5,7 +5,7 @@
 // http://opensource.org/licenses/MIT>, at your option. You may not use this file except in
 // accordance with one or both of these licenses.
 
-use std::collections::{BTreeMap, HashMap};
+use std::collections::{BTreeMap, HashMap, HashSet};
 use std::net::SocketAddr;
 use std::sync::atomic::{AtomicU32, Ordering};
 use std::sync::{Arc, Mutex, RwLock};
@@ -19,7 +19,7 @@ use bip157::{
 	SyncUpdate, TrustedPeer, Warning,
 };
 use bitcoin::constants::SUBSIDY_HALVING_INTERVAL;
-use bitcoin::{Amount, FeeRate, Network, Script, ScriptBuf, Transaction, Txid};
+use bitcoin::{Amount, FeeRate, Network, OutPoint, Script, ScriptBuf, Transaction, Txid};
 use electrum_client::ElectrumApi;
 use lightning::chain::{Confirm, WatchedOutput};
 use lightning::util::ser::Writeable;
@@ -705,6 +705,7 @@ impl CbfChainSource {
 		all_scripts.extend(self.registered_scripts.lock().unwrap().iter().cloned());
 		let skip_height =
 			onchain_wallet.latest_checkpoint().height().checked_sub(REORG_SAFETY_BLOCKS);
+		let script_set: HashSet<ScriptBuf> = all_scripts.iter().cloned().collect();
 		let (sync_update, matched) = self.run_filter_scan(all_scripts, skip_height).await?;
 
 		log_debug!(
@@ -713,10 +714,25 @@ impl CbfChainSource {
 			matched.len()
 		);
 
-		// Fetch matching blocks and include all their transactions.
-		// The compact block filter already matched our scripts (covering both
-		// created outputs and spent inputs), so we include every transaction
-		// from matched blocks and let BDK determine relevance.
+		// Fetch matching blocks and include only script-relevant transactions.
+		// The compact filter gives us block-level matches; here we narrow to
+		// individual transactions that create or spend outputs to watched
+		// scripts. This avoids pushing thousands of irrelevant txs into BDK's
+		// TxGraph on wide filter scans.
+
+		// Seed the known-outpoint set with outputs from BDK's existing tx
+		// graph that pay to our scripts, so we can also detect txs that
+		// *spend* our UTXOs (not just receive to them).
+		let mut known_outpoints: HashSet<OutPoint> = HashSet::new();
+		for cached_tx in onchain_wallet.get_cached_txs() {
+			let cached_txid = cached_tx.compute_txid();
+			for (vout, output) in cached_tx.output.iter().enumerate() {
+				if script_set.contains(&output.script_pubkey) {
+					known_outpoints.insert(OutPoint::new(cached_txid, vout as u32));
+				}
+			}
+		}
+
 		let mut tx_update = TxUpdate::default();
 		let per_request_timeout =
 			Duration::from_secs(self.sync_config.timeouts_config.per_request_timeout_secs.into());
@@ -737,9 +753,21 @@ impl CbfChainSource {
 			let conf_time =
 				ConfirmationBlockTime { block_id, confirmation_time: block.header.time as u64 };
 			for tx in &block.txdata {
-				let txid = tx.compute_txid();
-				tx_update.txs.push(Arc::new(tx.clone()));
-				tx_update.anchors.insert((conf_time, txid));
+				let dominated = tx.output.iter().any(|o| script_set.contains(&o.script_pubkey));
+				let spends_known =
+					tx.input.iter().any(|i| known_outpoints.contains(&i.previous_output));
+
+				if dominated || spends_known {
+					let txid = tx.compute_txid();
+					tx_update.txs.push(Arc::new(tx.clone()));
+					tx_update.anchors.insert((conf_time, txid));
+
+					for (vout, output) in tx.output.iter().enumerate() {
+						if script_set.contains(&output.script_pubkey) {
+							known_outpoints.insert(OutPoint::new(txid, vout as u32));
+						}
+					}
+				}
 			}
 		}