From 712a03a3a63157885079954aeba014240a00044e Mon Sep 17 00:00:00 2001
From: Eric Coissac <eric@coissac.eu>
Date: Thu, 4 Jun 2026 20:55:24 +0200
Subject: [PATCH] refactor: replace unitig extraction with de Bruijn graph
 pipeline

This change replaces direct partition-based extraction with a pipeline that reconstructs a de Bruijn graph from filtered k-mers. It introduces `FilterArgs` for k-mer selection, collects filtered k-mers in parallel into a `GraphDeBruijn`, computes node degrees, and enumerates unitigs from the graph for output instead of reading pre-computed partition files.
---
 src/obikmer/src/cmd/unitig.rs | 69 ++++++++++++++++++++---------------
 1 file changed, 40 insertions(+), 29 deletions(-)

diff --git a/src/obikmer/src/cmd/unitig.rs b/src/obikmer/src/cmd/unitig.rs
index 82fe6bc..6c36bb6 100644
--- a/src/obikmer/src/cmd/unitig.rs
+++ b/src/obikmer/src/cmd/unitig.rs
@@ -1,54 +1,65 @@
 use std::io::{self, BufWriter, Write};
 use std::path::PathBuf;
-use std::sync::Mutex;
 
 use clap::Args;
+use obidebruinj::GraphDeBruijn;
 use obifastwrite::write_unitig;
 use obikindex::KmerIndex;
-use obikseq::set_k;
-use obiskio::UnitigFileReader;
-use rayon::prelude::*;
 use tracing::info;
 
+use super::predicate::FilterArgs;
+
 #[derive(Args)]
 pub struct UnitigArgs {
-    /// Index directory produced by the `index` command
+    /// Index directory
     pub index: PathBuf,
+
+    #[command(flatten)]
+    pub filter: FilterArgs,
 }
 
 pub fn run(args: UnitigArgs) {
     let idx = KmerIndex::open(&args.index).unwrap_or_else(|e| {
         eprintln!("error opening index: {e}");
-        std::process::exit(1)
+        std::process::exit(1);
     });
 
-    let k = idx.kmer_size();
-    set_k(k);
-    let n = idx.n_partitions();
-    info!("dumping unitigs from {n} partitions (k={k})");
+    let k          = idx.kmer_size();
+    let n          = idx.n_partitions();
+    let n_genomes  = idx.meta().genomes.len().max(1);
+    let use_counts = idx.meta().config.with_counts;
 
-    let stdout = Mutex::new(BufWriter::new(io::stdout()));
+    info!("unitig: building de Bruijn graph from {n} partition(s) (k={k})");
 
-    (0..n).into_par_iter().for_each(|i| {
-        let path = idx.layer_unitigs_path(i, 0);
-        if !path.exists() {
-            return;
-        }
+    let filters = args.filter.build_filters(&idx.meta().genomes);
 
-        // open_sequential: works with and without .idx (approx or exact index)
-        let reader = UnitigFileReader::open_sequential(&path).unwrap_or_else(|e| {
-            eprintln!("error opening unitigs (partition {i}): {e}");
-            std::process::exit(1)
-        });
+    // ── Collect all filtered kmers into a single de Bruijn graph ─────────────
+    let mut g = GraphDeBruijn::new();
 
-        for (j, unitig) in reader.iter_unitigs() {
-            let mut out = stdout.lock().unwrap();
-            write_unitig(&unitig, k, i, j, &mut *out).unwrap_or_else(|e| {
-                eprintln!("write error: {e}");
-                std::process::exit(1)
+    for i in 0..n {
+        idx.partition()
+            .iter_partition_kmers(i, use_counts, n_genomes, &filters, |kmer, _row| {
+                g.push(kmer);
+            })
+            .unwrap_or_else(|e| {
+                eprintln!("error reading partition {i}: {e}");
+                std::process::exit(1);
             });
-        }
-    });
+    }
 
-    stdout.into_inner().unwrap().flush().expect("flush error");
+    info!("unitig: {} distinct k-mers", g.len());
+    g.compute_degrees();
+
+    // ── Enumerate unitigs and write as FASTA ──────────────────────────────────
+    let stdout = io::stdout();
+    let mut out = BufWriter::new(stdout.lock());
+
+    for (j, unitig) in g.iter_unitig().enumerate() {
+        write_unitig(&unitig, k, 0, j, &mut out).unwrap_or_else(|e| {
+            eprintln!("write error: {e}");
+            std::process::exit(1);
+        });
+    }
+
+    out.flush().expect("flush error");
 }