📖 Update super-kmer theory and implementation to prefer non-degenerate m-mers

- Update super-kmer definition in `kmERS.md` to specify that non-degenerate m-mers are preferred over degenerate ones (degeneracy = homopolymer).
- Refactor `superkmer.rs`: change `.canonical()` to mutate in-place and return bool.
- Add `m` field & canonical-aware minimizer position calculation to SuperKmerIter in obiskbuilder.
- Add helper functions `is_degenerate` and minimizer comparison logic to rolling_stat.rs for consistent tie-breaking.
- Minor formatting cleanup in superkmer command and chunk processing.

src/obikseq/src/superkmer.rs

@@ -286,21 +286,23 @@ impl SuperKmer {
         Ok(self.kmer(i, k)?.canonical(k))
     }
 
-    /// Return this super-kmer in canonical form (lexicographic minimum of forward and revcomp).
-    pub fn canonical(mut self) -> Self {
+    /// Put this super-kmer in canonical form (lexicographic minimum of forward and revcomp).
+    ///
+    /// Returns `true` if already canonical (no change), `false` if revcomp was applied.
+    pub fn canonical(&mut self) -> bool {
         let seql = self.seql();
         for i in 0..seql {
             let fwd = self.nucleotide(i);
             let rev = complement(self.nucleotide(seql - 1 - i));
             if fwd < rev {
-                return self;
+                return true;
             }
             if fwd > rev {
                 self.revcomp();
-                return self;
+                return false;
             }
         }
-        self
+        true
     }
 }