wip: updates to musals

benches/cakes/src/metric/cosine.rs

@@ -24,7 +24,7 @@ impl<I: AsRef<[f32]>> Metric<I, f32> for Cosine {
         distances::simd::cosine_f32(a.as_ref(), b.as_ref())
     }
 
-    fn name(&self) -> &str {
+    fn name(&self) -> &'static str {
         "cosine"
     }
 
@@ -89,7 +89,7 @@ impl<I: AsRef<[f64]>> Metric<I, f64> for Cosine {
         distances::simd::cosine_f64(a.as_ref(), b.as_ref())
     }
 
-    fn name(&self) -> &str {
+    fn name(&self) -> &'static str {
         "cosine"
     }
 

benches/cakes/src/metric/dtw.rs

@@ -25,7 +25,7 @@ impl<I: AsRef<[Complex<f64>]>> Metric<I, f64> for DynamicTimeWarping {
         dtw_distance(a.as_ref(), b.as_ref())
     }
 
-    fn name(&self) -> &str {
+    fn name(&self) -> &'static str {
         "dtw"
     }
 

benches/cakes/src/metric/euclidean.rs

@@ -24,7 +24,7 @@ impl<I: AsRef<[f32]>> Metric<I, f32> for Euclidean {
         distances::simd::euclidean_f32(a.as_ref(), b.as_ref())
     }
 
-    fn name(&self) -> &str {
+    fn name(&self) -> &'static str {
         "euclidean"
     }
 
@@ -89,7 +89,7 @@ impl<I: AsRef<[f64]>> Metric<I, f64> for Euclidean {
         distances::simd::euclidean_f64(a.as_ref(), b.as_ref())
     }
 
-    fn name(&self) -> &str {
+    fn name(&self) -> &'static str {
         "euclidean"
     }
 

benches/cakes/src/metric/hamming.rs

@@ -24,7 +24,7 @@ impl Metric<String, u32> for Hamming {
         distances::strings::hamming(a, b)
     }
 
-    fn name(&self) -> &str {
+    fn name(&self) -> &'static str {
         "hamming"
     }
 

benches/cakes/src/metric/levenshtein.rs

@@ -25,7 +25,7 @@ impl Metric<String, u32> for Levenshtein {
         stringzilla::sz::edit_distance(a, b).as_u32()
     }
 
-    fn name(&self) -> &str {
+    fn name(&self) -> &'static str {
         "levenshtein"
     }
 

benches/musals/src/main.rs

@@ -17,7 +17,7 @@
 
 use std::path::PathBuf;
 
-use abd_clam::{metric::Levenshtein, msa, Cluster, Dataset};
+use abd_clam::{metric::Levenshtein, musals, Cluster, Dataset};
 use bench_utils::configure_logger;
 use clap::Parser;
 use distances::Number;
@@ -39,8 +39,8 @@ struct Args {
     num_samples: Option<usize>,
 
     /// The cost matrix to use for the alignment.
-    #[arg(short('m'), long)]
-    cost_matrix: CostMatrix,
+    #[arg(short('m'), long, default_value = "default")]
+    matrix: CostMatrix,
 
     /// Optional cost of opening a gap, if using an affine gap penalty. If not
     /// provided, a flat gap penalty is used.
@@ -55,7 +55,8 @@ struct Args {
     #[arg(short('o'), long)]
     out_dir: Option<PathBuf>,
 
-    /// Whether to only compute the quality metrics.
+    /// Whether to only compute the quality metrics. This will assume that the
+    /// input file contains aligned sequences.
     #[arg(short('q'), long)]
     quality_only: bool,
 }
@@ -82,12 +83,12 @@ pub enum CostMatrix {
 impl CostMatrix {
     /// Get the cost matrix.
     #[must_use]
-    pub fn cost_matrix<T: Number + core::ops::Neg<Output = T>>(&self, gap_open: Option<usize>) -> msa::CostMatrix<T> {
+    pub fn cost_matrix<T: Number + core::ops::Neg<Output = T>>(&self, gap_open: Option<usize>) -> musals::CostMatrix<T> {
         match self {
-            Self::Default => msa::CostMatrix::default(),
-            Self::DefaultAffine => msa::CostMatrix::default_affine(gap_open),
-            Self::ExtendedIupac => msa::CostMatrix::extended_iupac(gap_open),
-            Self::Blosum62 => msa::CostMatrix::blosum62(gap_open),
+            Self::Default => musals::CostMatrix::default(),
+            Self::DefaultAffine => musals::CostMatrix::default_affine(gap_open),
+            Self::ExtendedIupac => musals::CostMatrix::extended_iupac(gap_open),
+            Self::Blosum62 => musals::CostMatrix::blosum62(gap_open),
         }
     }
 }
@@ -109,8 +110,8 @@ fn main() -> Result<(), String> {
     println!("Log file: {log_path:?}");
     ftlog::info!("{args:?}");
 
-    let cost_matrix = args.cost_matrix.cost_matrix::<i32>(args.gap_open);
-    let aligner = msa::Aligner::new(&cost_matrix, b'-');
+    let cost_matrix = args.matrix.cost_matrix::<i32>(args.gap_open);
+    let aligner = musals::Aligner::new(&cost_matrix, b'-');
 
     let out_dir = if let Some(out_dir) = args.out_dir {
         if !out_dir.exists() {
@@ -142,7 +143,7 @@ fn main() -> Result<(), String> {
         } else {
             data
         };
-        let str_to_seq = |s: String| msa::Sequence::new(s, Some(&aligner));
+        let str_to_seq = |s: String| musals::Sequence::new(s, Some(&aligner));
         let data = data.transform_items(str_to_seq);
         ftlog::info!(
             "Finished reading original dataset: length range = {:?}",
@@ -179,13 +180,7 @@ fn main() -> Result<(), String> {
             ftlog::info!("Permuted Ball has {} leaves.", perm_ball.leaves().len());
 
             // Build the MSA.
-            steps::build_aligned(
-                &args.cost_matrix,
-                args.gap_open,
-                &perm_ball,
-                &perm_data,
-                &msa_fasta_path,
-            )?;
+            steps::build_aligned(&args.matrix, args.gap_open, &perm_ball, &perm_data, &msa_fasta_path)?;
             let elapsed = start.elapsed().as_secs_f32();
             let msa_build_msg = format!("Finished building MSA in {elapsed:.2} seconds.");
             ftlog::info!("{msa_build_msg}");

benches/musals/src/steps.rs

@@ -7,7 +7,7 @@ use abd_clam::{
     cluster::{adapter::ParBallAdapter, BalancedBall, ClusterIO, Csv, ParPartition},
     dataset::{AssociatesMetadata, AssociatesMetadataMut, DatasetIO},
     metric::ParMetric,
-    msa::{self, Aligner, Sequence},
+    musals::{Aligner, Columns, Sequence, MSA},
     Ball, Cluster, Dataset, FlatVec,
 };
 
@@ -25,18 +25,18 @@ pub fn build_aligned<P: AsRef<Path>>(
     ftlog::info!("Setting up aligner...");
     let gap = b'-';
     let cost_matrix = matrix.cost_matrix::<i32>(gap_open);
-    let aligner = msa::Aligner::new(&cost_matrix, gap);
+    let aligner = Aligner::new(&cost_matrix, gap);
 
     ftlog::info!("Aligning sequences...");
-    let builder = msa::Columns::new(gap).par_with_tree(perm_ball, data, &aligner);
+    let builder = Columns::new(gap).par_with_tree(perm_ball, data, &aligner);
 
     ftlog::info!("Extracting aligned sequences...");
     let msa = builder.to_flat_vec_rows().with_metadata(data.metadata())?;
     let transformer = |s: Vec<u8>| s.into_iter().map(|c| c as char).collect::<String>();
     let msa = msa.transform_items(transformer);
 
     ftlog::info!("Finished aligning {} sequences.", builder.len());
-    let data = msa::MSA::new(&aligner, msa)?;
+    let data = MSA::new(&aligner, msa)?;
 
     ftlog::info!("Writing MSA to {:?}", out_path.as_ref());
     bench_utils::fasta::write(&data, out_path)?;
@@ -45,10 +45,10 @@ pub fn build_aligned<P: AsRef<Path>>(
 }
 
 /// Read the aligned fasta file.
-pub fn read_aligned<P: AsRef<Path>>(path: &P, aligner: &Aligner<i32>) -> Result<msa::MSA<String, i32, String>, String> {
+pub fn read_aligned<P: AsRef<Path>>(path: &P, aligner: &Aligner<i32>) -> Result<MSA<String, i32, String>, String> {
     ftlog::info!("Reading aligned sequences from {:?}", path.as_ref());
     let (data, _) = bench_utils::fasta::read(path, 0, false)?;
-    msa::MSA::new(aligner, data)
+    MSA::new(aligner, data)
 }
 
 /// Build the `PermutedBall` and the permuted dataset.

benches/utils/src/fasta/mod.rs

@@ -43,7 +43,7 @@ pub fn read<P: AsRef<Path>>(
         return Err(format!("Path {path:?} does not exist!"));
     }
 
-    if !path.extension().map_or(false, |ext| ext == "fasta") {
+    if path.extension().is_none_or(|ext| ext != "fasta") {
         return Err(format!("Path {path:?} does not have the `.fasta` extension!"));
     }
 

benches/utils/src/metrics/dtw.rs

@@ -61,7 +61,7 @@ impl<I: AsRef<[Complex<f32>]>> Metric<I, f32> for Dtw {
         dtw_distance(a.as_ref(), b.as_ref())
     }
 
-    fn name(&self) -> &str {
+    fn name(&self) -> &'static str {
         "dtw"
     }
 
@@ -93,7 +93,7 @@ impl<I: AsRef<[Complex<f64>]>> Metric<I, f64> for Dtw {
         dtw_distance(a.as_ref(), b.as_ref())
     }
 
-    fn name(&self) -> &str {
+    fn name(&self) -> &'static str {
         "dtw"
     }
 

benches/utils/src/metrics/jaccard.rs

@@ -10,7 +10,7 @@ impl<I: AsRef<[usize]>> Metric<I, f32> for Jaccard {
         distances::sets::jaccard(a.as_ref(), b.as_ref())
     }
 
-    fn name(&self) -> &str {
+    fn name(&self) -> &'static str {
         "jaccard"
     }
 

benches/utils/src/metrics/levenshtein.rs

@@ -11,7 +11,7 @@ impl<I: AsRef<[u8]>, T: Number> Metric<I, T> for Levenshtein {
         T::from(stringzilla::sz::edit_distance(a.as_ref(), b.as_ref()))
     }
 
-    fn name(&self) -> &str {
+    fn name(&self) -> &'static str {
         "levenshtein"
     }
 

crates/abd-clam/Cargo.toml

@@ -67,8 +67,8 @@ statistical = "1.0.0"
 disk-io = ["dep:serde", "dep:csv", "dep:bitcode", "dep:flate2", "dep:ndarray", "dep:ndarray-npy"]
 chaoda = ["dep:smartcore"]
 mbed = ["chaoda"]
-msa = ["dep:stringzilla", "dep:bytecount"]
-all = ["disk-io", "chaoda", "mbed", "msa"]
+musals = ["dep:stringzilla", "dep:bytecount"]
+all = ["disk-io", "chaoda", "mbed", "musals"]
 
 [[bench]]
 name = "vector_search"
@@ -77,7 +77,7 @@ harness = false
 [[bench]]
 name = "genomic_search"
 harness = false
-required-features = ["msa"]
+required-features = ["musals"]
 
 [[bench]]
 name = "ann_benchmarks"

crates/abd-clam/README.md

@@ -15,7 +15,7 @@ This crate provides the following features:
 
 - `disk-io`: Enables easy IO for several structs, primarily using `bitcode` and `serde`.
 - `chaoda`: Enables anomaly detection using the CHAODA.
-- `msa`: Enables multiple sequence alignment.
+- `musals`: Enables multiple sequence alignment.
 - `mbed`: Enables dimensionality reduction algorithms.
 - `all`: Enables all features.
 
@@ -103,7 +103,7 @@ use abd_clam::{
     cluster::{adapter::ParBallAdapter, ClusterIO, ParPartition},
     dataset::{AssociatesMetadataMut, DatasetIO},
     metric::Levenshtein,
-    msa::{Aligner, CostMatrix, Sequence},
+    musals::{Aligner, CostMatrix, Sequence},
     pancakes::{CodecData, SquishyBall},
     Ball, Cluster, Dataset, FlatVec,
 };

crates/abd-clam/benches/genomic_search.rs

@@ -7,7 +7,7 @@ use abd_clam::{
     cluster::{adapter::ParBallAdapter, BalancedBall, ParPartition},
     dataset::{AssociatesMetadata, AssociatesMetadataMut},
     metric::Levenshtein,
-    msa::{Aligner, CostMatrix, Sequence},
+    musals::{Aligner, CostMatrix, Sequence},
     pancakes::SquishyBall,
     Ball, Cluster, Dataset, FlatVec,
 };

crates/abd-clam/src/cakes/search/knn_breadth_first.rs

@@ -20,7 +20,7 @@ pub struct KnnBreadthFirst(pub usize);
 impl<I, T: Number, C: Cluster<T>, M: Metric<I, T>, D: Searchable<I, T, C, M>> SearchAlgorithm<I, T, C, M, D>
     for KnnBreadthFirst
 {
-    fn name(&self) -> &str {
+    fn name(&self) -> &'static str {
         "KnnBreadthFirst"
     }
 

crates/abd-clam/src/cakes/search/knn_depth_first.rs

@@ -21,7 +21,7 @@ pub struct KnnDepthFirst(pub usize);
 impl<I, T: Number, C: Cluster<T>, M: Metric<I, T>, D: Searchable<I, T, C, M>> SearchAlgorithm<I, T, C, M, D>
     for KnnDepthFirst
 {
-    fn name(&self) -> &str {
+    fn name(&self) -> &'static str {
         "KnnDepthFirst"
     }
 

crates/abd-clam/src/cakes/search/knn_hinted.rs

@@ -23,7 +23,7 @@ where
     M: Metric<I, T>,
     D: HintedDataset<I, T, C, M>,
 {
-    fn name(&self) -> &str {
+    fn name(&self) -> &'static str {
         "KnnHinted"
     }
 

crates/abd-clam/src/cakes/search/knn_linear.rs

@@ -18,7 +18,7 @@ pub struct KnnLinear(pub usize);
 impl<I, T: Number, C: Cluster<T>, M: Metric<I, T>, D: Searchable<I, T, C, M>> SearchAlgorithm<I, T, C, M, D>
     for KnnLinear
 {
-    fn name(&self) -> &str {
+    fn name(&self) -> &'static str {
         "KnnLinear"
     }
 

crates/abd-clam/src/cakes/search/knn_repeated_rnn.rs

@@ -21,7 +21,7 @@ pub struct KnnRepeatedRnn<T: Number>(pub usize, pub T);
 impl<I, T: Number, C: Cluster<T>, M: Metric<I, T>, D: Searchable<I, T, C, M>> SearchAlgorithm<I, T, C, M, D>
     for KnnRepeatedRnn<T>
 {
-    fn name(&self) -> &str {
+    fn name(&self) -> &'static str {
         "KnnRepeatedRnn"
     }
 

crates/abd-clam/src/cakes/search/rnn_clustered.rs

@@ -19,7 +19,7 @@ pub struct RnnClustered<T: Number>(pub T);
 impl<I, T: Number, C: Cluster<T>, M: Metric<I, T>, D: Searchable<I, T, C, M>> SearchAlgorithm<I, T, C, M, D>
     for RnnClustered<T>
 {
-    fn name(&self) -> &str {
+    fn name(&self) -> &'static str {
         "RnnClustered"
     }
 

crates/abd-clam/src/cakes/search/rnn_linear.rs

@@ -18,7 +18,7 @@ pub struct RnnLinear<T: Number>(pub T);
 impl<I, T: Number, C: Cluster<T>, M: Metric<I, T>, D: Searchable<I, T, C, M>> SearchAlgorithm<I, T, C, M, D>
     for RnnLinear<T>
 {
-    fn name(&self) -> &str {
+    fn name(&self) -> &'static str {
         "RnnLinear"
     }
 

crates/abd-clam/src/chaoda/training/algorithms/cc.rs

@@ -12,7 +12,7 @@ use super::GraphEvaluator;
 pub struct ClusterCardinality;
 
 impl<T: Number, S: Cluster<T>> GraphEvaluator<T, S> for ClusterCardinality {
-    fn name(&self) -> &str {
+    fn name(&self) -> &'static str {
         "cc"
     }
 

crates/abd-clam/src/chaoda/training/algorithms/gn.rs

@@ -30,7 +30,7 @@ impl GraphNeighborhood {
 }
 
 impl<T: Number, S: Cluster<T>> GraphEvaluator<T, S> for GraphNeighborhood {
-    fn name(&self) -> &str {
+    fn name(&self) -> &'static str {
         "gn"
     }
 

crates/abd-clam/src/chaoda/training/algorithms/pc.rs

@@ -13,7 +13,7 @@ use super::GraphEvaluator;
 pub struct ParentCardinality;
 
 impl<T: Number, S: Cluster<T>> GraphEvaluator<T, S> for ParentCardinality {
-    fn name(&self) -> &str {
+    fn name(&self) -> &'static str {
         "pc"
     }
 

crates/abd-clam/src/chaoda/training/algorithms/sc.rs

@@ -13,7 +13,7 @@ use super::GraphEvaluator;
 pub struct SubgraphCardinality;
 
 impl<T: Number, S: Cluster<T>> GraphEvaluator<T, S> for SubgraphCardinality {
-    fn name(&self) -> &str {
+    fn name(&self) -> &'static str {
         "sc"
     }
 

crates/abd-clam/src/chaoda/training/algorithms/sp.rs

@@ -26,7 +26,7 @@ impl StationaryProbability {
 }
 
 impl<T: Number, S: Cluster<T>> GraphEvaluator<T, S> for StationaryProbability {
-    fn name(&self) -> &str {
+    fn name(&self) -> &'static str {
         "sp"
     }
 

crates/abd-clam/src/chaoda/training/algorithms/vd.rs

@@ -12,7 +12,7 @@ use super::GraphEvaluator;
 pub struct VertexDegree;
 
 impl<T: Number, S: Cluster<T>> GraphEvaluator<T, S> for VertexDegree {
-    fn name(&self) -> &str {
+    fn name(&self) -> &'static str {
         "vd"
     }