WIP: skipmer improvements (#3415)

Make skipmers robust, but keep #3395 functional in the meantime.

This PR:
- enables second skipmer types, so we have m1n3 in addition to m2n3
- switches to a reading frame approach for both translation + skipmers,
which means we first build the reading frame, then kmerize, rather than
building kmers + translating/skipping on the fly
- avoids "extended length" needed for skipping on the fly

Since this changes the `SeqToHashes` strategy a bit, there's one python
test where we now see a different error.

Future thoughts:
- with the new structure, it would be straightforward to add validation
to exclude protein k-mers with invalid amino acids (`X`). I guess I'm
not entirely sure what happens to those atm...

src/core/src/cmd.rs

@@ -44,7 +44,11 @@ pub struct ComputeParameters {
 
     #[getset(get_copy = "pub", set = "pub")]
     #[builder(default = false)]
-    skipmer: bool,
+    skipm1n3: bool,
+
+    #[getset(get_copy = "pub", set = "pub")]
+    #[builder(default = false)]
+    skipm2n3: bool,
 
     #[getset(get_copy = "pub", set = "pub")]
     #[builder(default = false)]
@@ -169,12 +173,29 @@ pub fn build_template(params: &ComputeParameters) -> Vec<Sketch> {
                 ));
             }
 
-            if params.skipmer {
+            if params.skipm1n3 {
+                ksigs.push(Sketch::LargeMinHash(
+                    KmerMinHashBTree::builder()
+                        .num(params.num_hashes)
+                        .ksize(*k)
+                        .hash_function(HashFunctions::Murmur64Skipm1n3)
+                        .max_hash(max_hash)
+                        .seed(params.seed)
+                        .abunds(if params.track_abundance {
+                            Some(Default::default())
+                        } else {
+                            None
+                        })
+                        .build(),
+                ));
+            }
+
+            if params.skipm2n3 {
                 ksigs.push(Sketch::LargeMinHash(
                     KmerMinHashBTree::builder()
                         .num(params.num_hashes)
                         .ksize(*k)
-                        .hash_function(HashFunctions::Murmur64Skipmer)
+                        .hash_function(HashFunctions::Murmur64Skipm2n3)
                         .max_hash(max_hash)
                         .seed(params.seed)
                         .abunds(if params.track_abundance {

src/core/src/encodings.rs

@@ -31,7 +31,8 @@ pub enum HashFunctions {
     Murmur64Protein,
     Murmur64Dayhoff,
     Murmur64Hp,
-    Murmur64Skipmer,
+    Murmur64Skipm1n3,
+    Murmur64Skipm2n3,
     Custom(String),
 }
 
@@ -51,8 +52,13 @@ impl HashFunctions {
     pub fn hp(&self) -> bool {
         *self == HashFunctions::Murmur64Hp
     }
-    pub fn skipmer(&self) -> bool {
-        *self == HashFunctions::Murmur64Skipmer
+
+    pub fn skipm1n3(&self) -> bool {
+        *self == HashFunctions::Murmur64Skipm1n3
+    }
+
+    pub fn skipm2n3(&self) -> bool {
+        *self == HashFunctions::Murmur64Skipm2n3
     }
 }
 
@@ -66,7 +72,8 @@ impl std::fmt::Display for HashFunctions {
                 HashFunctions::Murmur64Protein => "protein",
                 HashFunctions::Murmur64Dayhoff => "dayhoff",
                 HashFunctions::Murmur64Hp => "hp",
-                HashFunctions::Murmur64Skipmer => "skipmer",
+                HashFunctions::Murmur64Skipm1n3 => "skipm1n3",
+                HashFunctions::Murmur64Skipm2n3 => "skipm2n3",
                 HashFunctions::Custom(v) => v,
             }
         )
@@ -82,7 +89,8 @@ impl TryFrom<&str> for HashFunctions {
             "dayhoff" => Ok(HashFunctions::Murmur64Dayhoff),
             "hp" => Ok(HashFunctions::Murmur64Hp),
             "protein" => Ok(HashFunctions::Murmur64Protein),
-            "skipmer" => Ok(HashFunctions::Murmur64Skipmer),
+            "skipm1n3" => Ok(HashFunctions::Murmur64Skipm1n3),
+            "skipm2n3" => Ok(HashFunctions::Murmur64Skipm2n3),
             v => unimplemented!("{v}"),
         }
     }