Extract index modules into specialized workspace subcrates

This commit partitions the obikindex crate into multiple focused subcrates (obikfilter, obikmerge, obikquery, obikrebuild, obikselect, obikstats, obikdump, and obikidxcache) to reduce coupling and clarify module boundaries. It standardizes error handling across the workspace using OKIError and OKIResult, updates index APIs to support lazy, disk-backed partition access, and migrates NUMA system utilities to a new obisys crate. All modifications are structural, focusing on dependency graph expansion, import path updates, and API surface reorganization without altering core runtime behavior.

src/obikdump/Cargo.toml

@@ -0,0 +1,8 @@
+[package]
+name = "obikdump"
+version = "0.1.0"
+edition = "2024"
+
+[dependencies]
+obikindex  = { path = "../obikindex" }
+obikfilter = { path = "../obikfilter" }

src/obikdump/src/dump.rs

@@ -0,0 +1,123 @@
+use std::io::Write;
+use std::sync::atomic::{AtomicUsize, Ordering};
+
+use rayon::prelude::*;
+
+use obikindex::{OKIError, OKIResult};
+use obikindex::KmerIndex;
+use obikfilter::KmerFilter;
+
+impl KmerIndex {
+    /// Write a CSV table of all indexed kmers to `out`.
+    ///
+    /// Columns: `kmer`, then one column per genome (in index order).
+    /// Values are counts (u32) when `use_counts = true`, otherwise 0/1.
+    ///
+    /// `force_presence` overrides `with_counts`: even if the index stores counts,
+    /// the output uses 0/1 presence columns.
+    ///
+    /// Partitions are scanned in parallel; each partition buffers its output locally
+    /// before the main thread writes the chunks in partition order.
+    ///
+    /// The caller must have set the global kmer length (`obikseq::set_k`) before
+    /// calling this method.
+    pub fn dump<W: Write, F: Fn() + Send + Sync>(
+        &self,
+        out: &mut W,
+        force_presence: bool,
+        debug: bool,
+        head: Option<usize>,
+        filters: &[Box<dyn KmerFilter>],
+        on_partition: F,
+    ) -> OKIResult<()> {
+        let genomes    = self.meta.genomes().map_err(OKIError::Io)?;
+        let use_counts = self.meta.config.with_counts && !force_presence;
+        let n_genomes  = genomes.len().max(1);
+        let kmer_size  = self.kmer_size();
+
+        // ── Header ────────────────────────────────────────────────────────────
+        if debug {
+            write!(out, "partition,layer,")?;
+        }
+        write!(out, "kmer")?;
+        for g in &genomes {
+            write!(out, ",{}", g.label)?;
+        }
+        writeln!(out)?;
+
+        // ── Rows — parallel over partitions ───────────────────────────────────
+        let n = self.n_partitions();
+
+        let write_row = |buf: &mut Vec<u8>, row: &[u32], prefix: &str| {
+            let _ = buf.write_all(prefix.as_bytes());
+            for &v in row { let _ = write!(buf, ",{v}"); }
+            let _ = buf.write_all(b"\n");
+        };
+
+        let chunks: Vec<OKIResult<Vec<u8>>> = if let Some(limit) = head {
+            // ── Bounded: atomic counter, early exit when limit reached ────────
+            let remaining = AtomicUsize::new(limit);
+            (0..n).into_par_iter().map(|i| {
+                if remaining.load(Ordering::Relaxed) == 0 { return Ok(vec![]); }
+                let mut buf = Vec::<u8>::new();
+                let try_write = |buf: &mut Vec<u8>, row: &[u32], prefix: &str| -> bool {
+                    match remaining.fetch_update(Ordering::SeqCst, Ordering::SeqCst, |cur| {
+                        if cur > 0 { Some(cur - 1) } else { None }
+                    }) {
+                        Err(_) => false,
+                        Ok(_)  => { write_row(buf, row, prefix); true }
+                    }
+                };
+                if debug {
+                    self
+                        .iter_partition_kmers_located(i, use_counts, n_genomes, filters, |part, layer, kmer, row| {
+                            let seq = String::from_utf8(kmer.to_ascii()).unwrap_or_else(|_| "?".repeat(kmer_size));
+                            try_write(&mut buf, &row, &format!("{part},{layer},{seq}"))
+                        })
+                        .map_err(OKIError::Partition)?;
+                } else {
+                    self
+                        .iter_partition_kmers(i, use_counts, n_genomes, filters, |kmer, row| {
+                            let seq = String::from_utf8(kmer.to_ascii()).unwrap_or_else(|_| "?".repeat(kmer_size));
+                            try_write(&mut buf, &row, &seq)
+                        })
+                        .map_err(OKIError::Partition)?;
+                }
+                on_partition();
+                Ok(buf)
+            }).collect()
+        } else {
+            // ── Unbounded: no atomic, no contention ───────────────────────────
+            (0..n).into_par_iter().map(|i| {
+                let mut buf = Vec::<u8>::new();
+                if debug {
+                    self
+                        .iter_partition_kmers_located(i, use_counts, n_genomes, filters, |part, layer, kmer, row| {
+                            let seq = String::from_utf8(kmer.to_ascii()).unwrap_or_else(|_| "?".repeat(kmer_size));
+                            write_row(&mut buf, &row, &format!("{part},{layer},{seq}"));
+                            true
+                        })
+                        .map_err(OKIError::Partition)?;
+                } else {
+                    self
+                        .iter_partition_kmers(i, use_counts, n_genomes, filters, |kmer, row| {
+                            let seq = String::from_utf8(kmer.to_ascii()).unwrap_or_else(|_| "?".repeat(kmer_size));
+                            write_row(&mut buf, &row, &seq);
+                            true
+                        })
+                        .map_err(OKIError::Partition)?;
+                }
+                on_partition();
+                Ok(buf)
+            }).collect()
+        };
+
+        // ── Sequential write ──────────────────────────────────────────────────
+        for chunk in chunks {
+            out.write_all(&chunk?)?;
+        }
+
+        out.flush()?;
+        Ok(())
+    }
+}

src/obikdump/src/lib.rs

@@ -0,0 +1,8 @@
+//! Raw content export of an `obikindex::KmerIndex`: staging ground for
+//! code currently living in `obikindex::index`'s dump path (`dump.rs`,
+//! `dump_layer.rs`), to be migrated here to lighten that crate. Kept as
+//! a separate crate — not a module of `obikindex` — so the dependency
+//! runs one way only (dump code depends on the data model, never the
+//! reverse), same pattern as `obikindexer`/`obikquery`.
+
+mod dump;