Skip to content
Draft
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
158 changes: 157 additions & 1 deletion src/core/benches/minhash.rs
Original file line number Diff line number Diff line change
Expand Up @@ -107,5 +107,161 @@ fn intersection(c: &mut Criterion) {
});
}

criterion_group!(minhash, intersection);
fn jaccard(c: &mut Criterion) {
let mut filename = PathBuf::from(env!("CARGO_MANIFEST_DIR"));
filename.push("../../tests/test-data/gather-abund/genome-s10.fa.gz.sig");
let file = File::open(filename).unwrap();
let reader = BufReader::new(file);
let mut sigs: Vec<Signature> = serde_json::from_reader(reader).expect("Loading error");
let mh = if let Sketch::MinHash(mh) = &sigs.swap_remove(0).sketches()[0] {
mh.clone()
} else {
unimplemented!()
};

let mut filename = PathBuf::from(env!("CARGO_MANIFEST_DIR"));
filename.push("../../tests/test-data/gather-abund/genome-s11.fa.gz.sig");
let file = File::open(filename).unwrap();
let reader = BufReader::new(file);
let mut sigs: Vec<Signature> = serde_json::from_reader(reader).expect("Loading error");
let mh2 = if let Sketch::MinHash(mh) = &sigs.swap_remove(0).sketches()[0] {
mh.clone()
} else {
unimplemented!()
};

let mut group = c.benchmark_group("minhash");
group.sample_size(10);

group.bench_function("jaccard", |b| {
b.iter(|| {
mh.jaccard(&mh2).unwrap();
});
});

let mut mh1 = KmerMinHash::builder()
.num(0)
.max_hash(1_000_000)
.ksize(21)
.build();
let mut mh2 = KmerMinHash::builder()
.num(0)
.max_hash(1_000_000)
.ksize(21)
.build();

let mut mh1_btree = KmerMinHashBTree::builder()
.num(0)
.max_hash(1_000_000)
.ksize(21)
.build();
let mut mh2_btree = KmerMinHashBTree::builder()
.num(0)
.max_hash(1_000_000)
.ksize(21)
.build();

for i in 0..=1_000_000 {
if i % 2 == 0 {
mh1.add_hash(i);
mh1_btree.add_hash(i);
}
if i % 45 == 0 {
mh2.add_hash(i);
mh2_btree.add_hash(i);
}
}

group.bench_function("large jaccard", |b| {
b.iter(|| {
mh1.jaccard(&mh2).unwrap();
});
});

group.bench_function("large jaccard btree", |b| {
b.iter(|| {
mh1_btree.jaccard(&mh2_btree).unwrap();
});
});
}

fn count_common(c: &mut Criterion) {
let mut filename = PathBuf::from(env!("CARGO_MANIFEST_DIR"));
filename.push("../../tests/test-data/gather-abund/genome-s10.fa.gz.sig");
let file = File::open(filename).unwrap();
let reader = BufReader::new(file);
let mut sigs: Vec<Signature> = serde_json::from_reader(reader).expect("Loading error");
let mh = if let Sketch::MinHash(mh) = &sigs.swap_remove(0).sketches()[0] {
mh.clone()
} else {
unimplemented!()
};

let mut filename = PathBuf::from(env!("CARGO_MANIFEST_DIR"));
filename.push("../../tests/test-data/gather-abund/genome-s11.fa.gz.sig");
let file = File::open(filename).unwrap();
let reader = BufReader::new(file);
let mut sigs: Vec<Signature> = serde_json::from_reader(reader).expect("Loading error");
let mh2 = if let Sketch::MinHash(mh) = &sigs.swap_remove(0).sketches()[0] {
mh.clone()
} else {
unimplemented!()
};

let mut group = c.benchmark_group("minhash");
group.sample_size(10);

group.bench_function("count_common", |b| {
b.iter(|| {
mh.count_common(&mh2, false).unwrap();
});
});

let mut mh1 = KmerMinHash::builder()
.num(0)
.max_hash(1_000_000)
.ksize(21)
.build();
let mut mh2 = KmerMinHash::builder()
.num(0)
.max_hash(1_000_000)
.ksize(21)
.build();

let mut mh1_btree = KmerMinHashBTree::builder()
.num(0)
.max_hash(1_000_000)
.ksize(21)
.build();
let mut mh2_btree = KmerMinHashBTree::builder()
.num(0)
.max_hash(1_000_000)
.ksize(21)
.build();

for i in 0..=1_000_000 {
if i % 2 == 0 {
mh1.add_hash(i);
mh1_btree.add_hash(i);
}
if i % 45 == 0 {
mh2.add_hash(i);
mh2_btree.add_hash(i);
}
}

group.bench_function("large count_common", |b| {
b.iter(|| {
mh1.count_common(&mh2, false).unwrap();
});
});

group.bench_function("large count_common btree", |b| {
b.iter(|| {
mh1_btree.count_common(&mh2_btree, false).unwrap();
});
});
}

criterion_group!(minhash, intersection, jaccard, count_common);
criterion_main!(minhash);
5 changes: 3 additions & 2 deletions src/core/src/ffi/minhash.rs
Original file line number Diff line number Diff line change
Expand Up @@ -8,6 +8,7 @@ use crate::ffi::HashFunctions;
use crate::signature::SeqToHashes;
use crate::signature::SigsTrait;
use crate::sketch::minhash::KmerMinHash;
use crate::HashIntoType;

pub struct SourmashKmerMinHash;

Expand Down Expand Up @@ -206,7 +207,7 @@ pub unsafe extern "C" fn kmerminhash_remove_many(
ffi_fn! {
unsafe fn kmerminhash_get_mins(ptr: *const SourmashKmerMinHash, size: *mut usize) -> Result<*const u64> {
let mh = SourmashKmerMinHash::as_rust(ptr);
let output = mh.mins();
let output: Vec<HashIntoType> = mh.iter_mins().copied().collect();
*size = output.len();

// FIXME: make a SourmashSlice_u64 type?
Expand Down Expand Up @@ -434,7 +435,7 @@ unsafe fn kmerminhash_intersection(ptr: *const SourmashKmerMinHash, other: *cons
let isect = mh.intersection(other_mh)?;
let mut new_mh = mh.clone();
new_mh.clear();
new_mh.add_many(&isect.0)?;
new_mh.add_many(isect.0.iter().copied())?;

Ok(SourmashKmerMinHash::from_rust(new_mh))
}
Expand Down
2 changes: 1 addition & 1 deletion src/core/src/ffi/signature.rs
Original file line number Diff line number Diff line change
Expand Up @@ -202,7 +202,7 @@ ffi_fn! {
unsafe fn signature_get_mhs(ptr: *const SourmashSignature, size: *mut usize) -> Result<*mut *mut SourmashKmerMinHash> {
let sig = SourmashSignature::as_rust(ptr);

let output = sig.sketches();
let output = sig.iter();

// FIXME: how to fit this into the ForeignObject trait?
let ptr_sigs: Vec<*mut Signature> = output.into_iter().map(|x| {
Expand Down
21 changes: 9 additions & 12 deletions src/core/src/from.rs
Original file line number Diff line number Diff line change
Expand Up @@ -16,7 +16,7 @@ impl From<MashSketcher> for KmerMinHash {

let mut new_mh = KmerMinHash::new(
0,
values.get(0).unwrap().kmer.len() as u32,
values.first().unwrap().kmer.len() as u32,
HashFunctions::Murmur64Dna,
42,
true,
Expand Down Expand Up @@ -65,16 +65,15 @@ mod test {

let b_hashes = b.to_vec();

let s1: HashSet<_> = a.mins().into_iter().collect();
let s1: HashSet<_> = a.iter_mins().copied().collect();
let s2: HashSet<_> = b_hashes.iter().map(|x| x.hash).collect();
let i1 = &s1 & &s2;

assert!(i1.len() == a.size());
assert!(i1.len() == b_hashes.len());

if let Some(abunds) = a.abunds() {
let mins = a.mins();
let smap: HashMap<_, _> = mins.iter().zip(abunds.iter()).collect();
let smap: HashMap<_, _> = a.iter_mins().zip(abunds.iter()).collect();
println!("{:?}", smap);
for item in b_hashes.iter() {
assert!(smap.contains_key(&{ item.hash }));
Expand All @@ -101,19 +100,17 @@ mod test {

let c = KmerMinHash::from(b);

let s1: HashSet<_> = a.mins().into_iter().collect();
let s2: HashSet<_> = c.mins().into_iter().collect();
let s1: HashSet<_> = a.iter_mins().collect();
let s2: HashSet<_> = c.iter_mins().collect();
let i1 = &s1 & &s2;

assert!(i1.len() == a.mins().len());
assert!(i1.len() == c.mins().len());
assert!(i1.len() == a.iter_mins().count());
assert!(i1.len() == c.iter_mins().count());

if let Some(a_abunds) = a.abunds() {
if let Some(c_abunds) = c.abunds() {
let a_mins = a.mins();
let a_smap: HashMap<_, _> = a_mins.iter().zip(a_abunds.iter()).collect();
let c_mins = c.mins();
let c_smap: HashMap<_, _> = c_mins.iter().zip(c_abunds.iter()).collect();
let a_smap: HashMap<_, _> = a.iter_mins().zip(a_abunds.iter()).collect();
let c_smap: HashMap<_, _> = c.iter_mins().zip(c_abunds.iter()).collect();
for item in a_smap.iter() {
assert!(c_smap.contains_key(*item.0));
assert!(c_smap.get(*item.0).unwrap() == item.1);
Expand Down
2 changes: 1 addition & 1 deletion src/core/src/index/linear.rs
Original file line number Diff line number Diff line change
Expand Up @@ -25,7 +25,7 @@ pub struct LinearIndex {
impl LinearIndex {
pub fn from_collection(collection: CollectionSet) -> Self {
let sig = collection.sig_for_dataset(0).unwrap();
let template = sig.sketches().swap_remove(0);
let template = sig.iter().next().unwrap().clone();
Self {
collection,
template,
Expand Down
8 changes: 4 additions & 4 deletions src/core/src/index/revindex/disk_revindex.rs
Original file line number Diff line number Diff line change
Expand Up @@ -191,15 +191,15 @@ impl RevIndex {
.collection
.sig_for_dataset(dataset_id)
.expect("Couldn't find a compatible Signature");
let search_mh = &search_sig.sketches()[0];
let search_mh = search_sig.iter().next().unwrap();

let colors = Datasets::new(&[dataset_id]).as_bytes().unwrap();

let cf_hashes = self.db.cf_handle(HASHES).unwrap();

let hashes = match search_mh {
Sketch::MinHash(mh) => mh.mins(),
Sketch::LargeMinHash(mh) => mh.mins(),
let hashes: Vec<_> = match search_mh {
Sketch::MinHash(mh) => mh.iter_mins().copied().collect(),
Sketch::LargeMinHash(mh) => mh.iter_mins().copied().collect(),
_ => unimplemented!(),
};

Expand Down
2 changes: 1 addition & 1 deletion src/core/src/index/revindex/mem_revindex.rs
Original file line number Diff line number Diff line change
Expand Up @@ -175,7 +175,7 @@ impl RevIndex {
}
}
} else {
let matched = search_mh.mins();
let matched: Vec<_> = search_mh.iter_mins().copied().collect();
let size = matched.len() as u64;
if !matched.is_empty() || size > threshold as u64 {
hash_to_color.add_to(&mut colors, dataset_id, matched);
Expand Down
9 changes: 1 addition & 8 deletions src/core/src/index/revindex/mod.rs
Original file line number Diff line number Diff line change
Expand Up @@ -19,7 +19,6 @@ use crate::index::{GatherResult, SigCounter};
use crate::prelude::*;
use crate::signature::Signature;
use crate::sketch::minhash::KmerMinHash;
use crate::sketch::Sketch;
use crate::HashIntoType;
use crate::Result;

Expand Down Expand Up @@ -225,13 +224,7 @@ impl RevIndex {
pub fn prepare_query(search_sig: Signature, selection: &Selection) -> Option<KmerMinHash> {
let sig = search_sig.select(selection).ok();

sig.and_then(|sig| {
if let Sketch::MinHash(mh) = sig.sketches().swap_remove(0) {
Some(mh)
} else {
None
}
})
sig.and_then(|sig| sig.minhash().cloned())
}

#[derive(Debug, Default, Clone)]
Expand Down
4 changes: 0 additions & 4 deletions src/core/src/signature.rs
Original file line number Diff line number Diff line change
Expand Up @@ -475,10 +475,6 @@ impl Signature {
self.signatures.len()
}

pub fn sketches(&self) -> Vec<Sketch> {
self.signatures.clone()
}

pub fn reset_sketches(&mut self) {
self.signatures = vec![];
}
Expand Down
4 changes: 2 additions & 2 deletions src/core/src/sketch/hyperloglog/mod.rs
Original file line number Diff line number Diff line change
Expand Up @@ -214,8 +214,8 @@ impl SigsTrait for HyperLogLog {

impl Update<HyperLogLog> for KmerMinHash {
fn update(&self, other: &mut HyperLogLog) -> Result<(), Error> {
for h in self.mins() {
other.add_hash(h);
for h in self.iter_mins() {
other.add_hash(*h);
}
Ok(())
}
Expand Down
Loading