MB-61640: Fuzzy Dynamic Scoring (#20)

Co-authored-by: Abhinav Dangeti <[email protected]>

.github/workflows/tests.yml

@@ -8,7 +8,7 @@ jobs:
   test:
     strategy:
       matrix:
-        go-version: [1.17.x, 1.18.x, 1.19.x]
+        go-version: [1.20.x, 1.21.x, 1.22.x]
         platform: [ubuntu-latest, macos-latest, windows-latest]
     runs-on: ${{ matrix.platform }}
     steps:

automaton.go

@@ -83,3 +83,8 @@ func (m *AlwaysMatch) Accept(int, byte) int {
 
 // creating an alwaysMatchAutomaton to avoid unnecessary repeated allocations.
 var alwaysMatchAutomaton = &AlwaysMatch{}
+
+type FuzzyAutomaton interface {
+	Automaton
+	EditDistance(int) uint8
+}

fst_iterator.go

@@ -44,6 +44,11 @@ type Iterator interface {
 	Close() error
 }
 
+type FuzzyIterator interface {
+	Iterator
+	EditDistance() uint8
+}
+
 // FSTIterator is a structure for iterating key/value pairs in this FST in
 // lexicographic order.  Iterators should be constructed with the FSTIterator
 // method on the parent FST structure.
@@ -61,6 +66,8 @@ type FSTIterator struct {
 	autStatesStack []int
 
 	nextStart []byte
+
+	editDistance uint8
 }
 
 func newIterator(f *FST, startKeyInclusive, endKeyExclusive []byte,
@@ -74,6 +81,10 @@ func newIterator(f *FST, startKeyInclusive, endKeyExclusive []byte,
 	return rv, nil
 }
 
+func (i *FSTIterator) EditDistance() uint8 {
+	return i.editDistance
+}
+
 // Reset resets the Iterator' internal state to allow for iterator
 // reuse (e.g. pooling).
 func (i *FSTIterator) Reset(f *FST,
@@ -206,6 +217,9 @@ OUTER:
 
 			cmp := bytes.Compare(i.keysStack, i.nextStart)
 			if cmp > 0 {
+				if fa, ok := i.aut.(FuzzyAutomaton); ok {
+					i.editDistance = fa.EditDistance(autCurr)
+				}
 				// in final state greater than start key
 				return nil
 			}

go.mod

@@ -1,6 +1,6 @@
 module github.com/blevesearch/vellum
 
-go 1.18
+go 1.21
 
 require (
 	github.com/bits-and-blooms/bitset v1.2.0

levenshtein/dfa.go

@@ -28,23 +28,27 @@ type DFA struct {
 	ed          uint8
 }
 
-/// Returns the initial state
+// Returns the initial state
 func (d *DFA) initialState() int {
 	return d.initState
 }
 
-/// Returns the Levenshtein distance associated to the
-/// current state.
+// Returns the Levenshtein distance associated to the
+// current state.
 func (d *DFA) distance(stateId int) Distance {
 	return d.distances[stateId]
 }
 
-/// Returns the number of states in the `DFA`.
+func (d *DFA) EditDistance(stateId int) uint8 {
+	return d.distances[stateId].distance()
+}
+
+// Returns the number of states in the `DFA`.
 func (d *DFA) numStates() int {
 	return len(d.transitions)
 }
 
-/// Returns the destination state reached after consuming a given byte.
+// Returns the destination state reached after consuming a given byte.
 func (d *DFA) transition(fromState int, b uint8) int {
 	return int(d.transitions[fromState][b])
 }