[MRG] Produce list of hashes from a sequence (#1653)

* test for seq_to_hashes

* working seq_to_hashes

* Update src/core/src/signature.rs

Co-authored-by: Luiz Irber <luizirber@users.noreply.github.com>

* Update src/core/src/signature.rs

Co-authored-by: Luiz Irber <luizirber@users.noreply.github.com>

* Update src/core/src/ffi/minhash.rs

Co-authored-by: Luiz Irber <luizirber@users.noreply.github.com>

* Update src/core/src/ffi/minhash.rs

Co-authored-by: Luiz Irber <luizirber@users.noreply.github.com>

* fixes to adapt luiz's suggestions

* support is_protein in seq_to_hashes

* extended tests for seq_to_hashes

* Update src/sourmash/minhash.py

Co-authored-by: C. Titus Brown <titus@idyll.org>

* SeqToHashes Iterator dna

* Update src/sourmash/minhash.py

Co-authored-by: C. Titus Brown <titus@idyll.org>

* iterator implemented, missing exceptions

* better performance?

* one step away

* python tests pass

* test for long seqs with all kmers are invalid

* dev comments

* ✔️ 🎉 all passes

* 🐛 fix adding a single protein kmer

* fix test for moltype

* retain protein kmers hashes order

* seq_to_hashes rust test case

* more rust test coverage

* iterator refactoring

* partially enhance add_protein

* fully enhance add_protein

* ignore ffi coverage

* final refactoring

Co-authored-by: Luiz Irber <luizirber@users.noreply.github.com>
Co-authored-by: C. Titus Brown <titus@idyll.org>

codecov.yml

@@ -1,6 +1,8 @@
-# config file for codecov, https://app.codecov.io/gh/dib-lab/sourmash/
+# config file for codecov, https://app.codecov.io/gh/sourmash-bio/sourmash/
 
 # use path fixing to properly report code coverage on source code
 # per https://docs.codecov.io/docs/fixing-paths
 fixes:
 - "sourmash::src/sourmash"
+ignores:
+- "src/core/src/ffi"

include/sourmash.h

@@ -235,6 +235,13 @@ void kmerminhash_remove_many(SourmashKmerMinHash *ptr,
 
 uint64_t kmerminhash_seed(const SourmashKmerMinHash *ptr);
 
+const uint64_t *kmerminhash_seq_to_hashes(SourmashKmerMinHash *ptr,
+                                          const char *sequence,
+                                          uintptr_t insize,
+                                          bool force,
+                                          bool is_protein,
+                                          uintptr_t *size);
+
 void kmerminhash_set_abundances(SourmashKmerMinHash *ptr,
                                 const uint64_t *hashes_ptr,
                                 const uint64_t *abunds_ptr,

src/core/src/ffi/minhash.rs

@@ -4,6 +4,7 @@ use std::slice;
 
 use crate::encodings::{aa_to_dayhoff, aa_to_hp, translate_codon, HashFunctions};
 use crate::ffi::utils::{ForeignObject, SourmashStr};
+use crate::signature::SeqToHashes;
 use crate::signature::SigsTrait;
 use crate::sketch::minhash::KmerMinHash;
 
@@ -58,6 +59,34 @@ unsafe fn kmerminhash_add_sequence(ptr: *mut SourmashKmerMinHash, sequence: *con
 }
 }
 
+ffi_fn! {
+unsafe fn kmerminhash_seq_to_hashes(ptr: *mut SourmashKmerMinHash, sequence: *const c_char, insize: usize, force: bool, is_protein: bool, size: *mut usize) ->
+Result<*const u64> {
+
+    let mh = SourmashKmerMinHash::as_rust_mut(ptr);
+
+    let buf = {
+        assert!(!ptr.is_null());
+        slice::from_raw_parts(sequence as *const u8, insize)
+    };
+
+    let mut output: Vec<u64> = Vec::with_capacity(insize);
+
+    for hash_value in SeqToHashes::new(buf, mh.ksize(), force, is_protein, mh.hash_function(), mh.seed()){
+        match hash_value{
+            Ok(0) => continue,
+            Ok(x) => output.push(x),
+            Err(err) => return Err(err),
+        }
+    }
+
+    *size = output.len();
+
+    // FIXME: make a SourmashSlice_u64 type?
+    Ok(Box::into_raw(output.into_boxed_slice()) as *const u64)
+}
+}
+
 ffi_fn! {
 unsafe fn kmerminhash_add_protein(ptr: *mut SourmashKmerMinHash, sequence: *const c_char) ->
     Result<()> {