From 2bc189e96259c3c1d3c7abe0b4e1cfcc5da47e9e Mon Sep 17 00:00:00 2001
From: Eric Coissac <eric@coissac.eu>
Date: Fri, 12 Jun 2026 16:28:03 +0200
Subject: [PATCH] feat: dynamically compute seed expansion based on RSS

Introduce a `peak_rss_bytes()` utility for accurate per-phase RAM measurement. Replace the genome-length heuristic with a dynamic seed expansion ratio based on actual RSS delta. Explicitly drop the `GraphDeBruijn` instance before MPHF construction to prevent resource contention and ensure proper memory management.
---
 src/obikindex/src/merge.rs              | 12 ++++++++----
 src/obikpartitionner/src/merge_layer.rs |  6 ++++--
 src/obilayeredmap/src/map.rs            |  1 +
 src/obisys/src/lib.rs                   |  6 ++++++
 4 files changed, 19 insertions(+), 6 deletions(-)

diff --git a/src/obikindex/src/merge.rs b/src/obikindex/src/merge.rs
index 0857013..f7879f4 100644
--- a/src/obikindex/src/merge.rs
+++ b/src/obikindex/src/merge.rs
@@ -5,7 +5,7 @@ use std::path::Path;
 use std::sync::atomic::{AtomicU64, Ordering};
 use std::sync::{Arc, Mutex};
 
-use obisys::{MemoryBudget, Reporter, Stage, available_memory_bytes, progress_bar, spinner};
+use obisys::{MemoryBudget, Reporter, Stage, available_memory_bytes, peak_rss_bytes, progress_bar, spinner};
 use rayon::prelude::*;
 use tracing::{debug, info};
 
@@ -200,21 +200,25 @@ impl KmerIndex {
             let worst_id    = order[0];
             let worst_bytes = partition_sizes[worst_id];
 
+            let rss_before_pilot = peak_rss_bytes();
             let worst_g_len = dst_partition
                 .merge_partition(worst_id, &srcs, mode, n_dst_genomes, block_bits, &evidence)
                 .map_err(OKIError::Partition)?;
+            let rss_after_pilot = peak_rss_bytes();
             pb.inc(1);
 
-            let seed_expansion = if worst_bytes > 0 {
-                worst_g_len as u64 * 16 * 1000 / worst_bytes
+            let pilot_rss = rss_after_pilot.saturating_sub(rss_before_pilot);
+            let seed_expansion = if worst_bytes > 0 && pilot_rss > 0 {
+                pilot_rss * 1000 / worst_bytes
             } else {
                 FALLBACK_EXPANSION
             };
 
             info!(
                 "merge_partitions: pilot partition {} — {} unitig bytes → {} new kmers, \
-                 expansion {:.2}×",
+                 RSS delta {}, expansion {:.2}×",
                 worst_id, worst_bytes, worst_g_len,
+                fmt_bytes(pilot_rss),
                 seed_expansion as f64 / 1000.0,
             );
 
diff --git a/src/obikpartitionner/src/merge_layer.rs b/src/obikpartitionner/src/merge_layer.rs
index 6b55f5d..3cc5a47 100644
--- a/src/obikpartitionner/src/merge_layer.rs
+++ b/src/obikpartitionner/src/merge_layer.rs
@@ -223,12 +223,14 @@ impl KmerPartition {
                 uw.write(&unitig)
             })?;
             uw.close()?;
+            let n = g.len();
+            drop(g); // release GraphDeBruijn before MPHF build
             Layer::<()>::build(&new_layer_dir, block_bits, evidence).map_err(olm_to_sk)?;
-            g.len()
+            n
         } else {
+            drop(g);
             0
         };
-        drop(g);
 
         let new_mphf = if any_new {
             Some(MphfOnly::open(&new_layer_dir).map_err(olm_to_sk)?)
diff --git a/src/obilayeredmap/src/map.rs b/src/obilayeredmap/src/map.rs
index 18d3c55..31cf20c 100644
--- a/src/obilayeredmap/src/map.rs
+++ b/src/obilayeredmap/src/map.rs
@@ -104,3 +104,4 @@ fn layer_dir(root: &Path, i: usize) -> PathBuf {
 #[cfg(test)]
 #[path = "tests/map.rs"]
 mod tests;
+
diff --git a/src/obisys/src/lib.rs b/src/obisys/src/lib.rs
index bf2e678..5f8a3d3 100644
--- a/src/obisys/src/lib.rs
+++ b/src/obisys/src/lib.rs
@@ -119,6 +119,12 @@ use sysinfo::System;
 ///
 /// On macOS, `available_memory()` can return 0 when the memory compressor
 /// inflates the page count; in that case we fall back to half of total memory.
+/// Returns the process peak RSS (high-water mark since process start).
+/// Monotonically increasing — use delta before/after a phase to measure its RAM cost.
+pub fn peak_rss_bytes() -> u64 {
+    rss_to_bytes(&get_rusage())
+}
+
 pub fn available_memory_bytes() -> u64 {
     let sys = System::new_all();
     let host_avail = match sys.available_memory() {