sort valiant actually faster than seqvaliant for small alphabets

breandan · Oct 2, 2023 · 76591ae · 76591ae
1 parent 0ce0fb8
commit 76591ae
Show file tree

Hide file tree

Showing 2 changed files with 15 additions and 15 deletions.
diff --git a/src/commonMain/kotlin/ai/hypergraph/kaliningraph/parsing/SortValiant.kt b/src/commonMain/kotlin/ai/hypergraph/kaliningraph/parsing/SortValiant.kt
@@ -49,7 +49,7 @@ fun CFG.sortwiseAlgebra(metric: ChoiceMetric): Ring<Sort> =
     times = { x, y -> join(x, y, metric) },
   )
 
-const val MAX_CAPACITY = 100
+var MAX_SORT_CAPACITY = 1000
 // X ⊗ Z := { w | <x, z> ∈ X × Z, (w -> xz) ∈ P }
 // Greedily selects candidate string fragments according to ChoiceMetric
 fun CFG.join(X: Sort, Z: Sort, metric: ChoiceMetric = { it.weight }): Sort =
@@ -66,7 +66,7 @@ fun CFG.join(X: Sort, Z: Sort, metric: ChoiceMetric = { it.weight }): Sort =
     val list = acc ?: mutableListOf()
     val idx = list.binarySearch(choice, Choice.comparator)
     if (idx < 0) list.add(-idx - 1, choice) // Only if not already present
-    list.apply { if (MAX_CAPACITY < size) removeLast() }
+    list.apply { if (MAX_SORT_CAPACITY < size) removeLast() }
   }.mapValues { it.value.toSet() }
 
 fun union(l: Sort, r: Sort): Sort =
@@ -94,7 +94,7 @@ data class Choice(val tokens: List<Σᐩ>, val weight: Float): Comparable<Choice
   override fun compareTo(other: Choice): Int = comparator.compare(this, other)
 
   operator fun plus(other: Choice) =
-    Choice(tokens + other.tokens, weight + other.weight)
+    Choice(sanitized + other.sanitized, weight + other.weight)
 
   val sanitized by lazy { tokens.filter { "ε" !in it } }
   val asString by lazy { sanitized.joinToString(" ") }

diff --git a/src/commonTest/kotlin/ai/hypergraph/kaliningraph/parsing/SetValiantTest.kt b/src/commonTest/kotlin/ai/hypergraph/kaliningraph/parsing/SetValiantTest.kt
@@ -360,7 +360,7 @@ class SetValiantTest {
     val leaves = tree.contents()
     assertEquals(expr, leaves)
 
-    val holExpr = "_ _ _ _ _ _ _ _ _ _"
+    val holExpr = "_ _ _ _"
 
     measureTime {
       val solutions = ocamlCFG.solve(holExpr, levMetric("( false curry )"))
@@ -864,19 +864,19 @@ Yield_Arg -> From_Keyword Test | Testlist_Endcomma
     println("Grammar size: ${bhcfg.size}")
     println("Solutions:")
 
-    val template = "_ _ _ _ _ _ _"
-    measureTime {
-      template.tokenizeByWhitespace()
-        .solve(bhcfg, fillers = bhcfg.terminals)
-        .map { it.replace("ε", "").tokenizeByWhitespace().joinToString(" ") }
-        .distinct()
-        .onEach { println(it) }
-        .toList().also { println("Found ${it.size} solutions.") }
-    }.also { println("Brute force solver took: ${it.inWholeMilliseconds}ms") }
+    val template = List(53) { "_" }.joinToString(" ")
+//    measureTime {
+//      template.tokenizeByWhitespace()
+//        .solve(bhcfg, fillers = bhcfg.terminals)
+//        .map { it.replace("ε", "").tokenizeByWhitespace().joinToString(" ") }
+//        .distinct()
+//        .onEach { println(it) }
+//        .toList().also { println("Found ${it.size} solutions.") }
+//    }.also { println("Brute force solver took: ${it.inWholeMilliseconds}ms") }
 
     measureTime {
-//      bhcfg.solve(template, {it.weight})
-      bhcfg.solveSeq(template)
+      bhcfg.solve(template, {it.weight})
+//      bhcfg.solveSeq(template)
         .onEach { assertTrue { it in bhcfg.language }; println(it) }
         .toList().also { println("Found ${it.size} solutions.") }
     }.also { println("Sequential solver took: ${it.inWholeMilliseconds}ms") }