add Bar-Hillel parser

breandan · Oct 13, 2023 · bd5ad7e · bd5ad7e
1 parent f798ffd
commit bd5ad7e
Show file tree

Hide file tree

Showing 4 changed files with 86 additions and 146 deletions.
diff --git a/src/commonMain/kotlin/ai/hypergraph/kaliningraph/parsing/BarHillel.kt b/src/commonMain/kotlin/ai/hypergraph/kaliningraph/parsing/BarHillel.kt
@@ -27,7 +27,7 @@ infix fun CFG.intersectLevFSA(fsa: FSA): CFG {
       second.coords().dominates(third.coords())
   }
 
-  // For each production A → BC in P , for every p, q, r ∈ Q,
+  // For each production A → BC in P, for every p, q, r ∈ Q,
   // we have the production [p,A,r] → [p,B,q] [q,C,r] in P′.
   val binaryProds =
     nonterminalProductions.map {
@@ -56,4 +56,42 @@ infix fun CFG.intersectLevFSA(fsa: FSA): CFG {
     .also { println("∩-grammar construction took: ${clock.elapsedNow().inWholeMilliseconds}ms") }
 //    .also { println(it.pretty) }
 //    .also { println(it.size) }
+}
+
+
+infix fun FSA.intersect(cfg: CFG) = cfg.intersect(this)
+
+infix fun CFG.intersect(fsa: FSA): CFG {
+  val clock = TimeSource.Monotonic.markNow()
+  val initFinal =
+    (fsa.init * fsa.final).map { (q, r) -> "START -> [$q,START,$r]" }
+
+  val transits =
+    fsa.Q.map { (q, a, r) -> "[$q,$a,$r] -> $a" }
+
+  // For each production A → BC in P, for every p, q, r ∈ Q,
+  // we have the production [p,A,r] → [p,B,q] [q,C,r] in P′.
+  val binaryProds =
+    nonterminalProductions.map {
+      val triples = fsa.states * fsa.states * fsa.states
+      val (A, B, C) = it.π1 to it.π2[0] to it.π2[1]
+      triples.map { (p, q, r) -> "[$p,$A,$r] -> [$p,$B,$q] [$q,$C,$r]" }
+    }.flatten()
+
+  // For every production A → σ in P, for every (p, σ, q) ∈ Q × Σ × Q
+  // such that δ(p, σ) = q we have the production [p, A, q] → σ in P′.
+  val unitProds =
+    unitProductions.map { (A, rhs) ->
+      val relevantTransits = fsa.Q.filter { it.π2 == rhs[0] }
+      relevantTransits.map { (p, σ, q) -> "[$p,$A,$q] -> $σ" }
+    }.flatten()
+
+  return (initFinal + transits + binaryProds + unitProds).joinToString("\n")
+    .parseCFG(normalize = false)
+    .also { println("∩-grammar has ${it.size} total productions") }
+    .removeVestigalProductions().normalForm.noNonterminalStubs
+    .also { println("∩-grammar has ${it.size} useful productions") }
+    .also { println("∩-grammar construction took: ${clock.elapsedNow().inWholeMilliseconds}ms") }
+//    .also { println(it.pretty) }
+//    .also { println(it.size) }
 }
diff --git a/src/commonMain/kotlin/ai/hypergraph/kaliningraph/parsing/FSA.kt b/src/commonMain/kotlin/ai/hypergraph/kaliningraph/parsing/FSA.kt
@@ -28,4 +28,34 @@ data class FSA(val Q: TSA, val init: Set<Σᐩ>, val final: Set<Σᐩ>) {
     } intersect final).isNotEmpty()
 }
 
-val TSA.states by cache { flatMap { listOf(it.π1, it.π3) }.toSet() }
+val TSA.states by cache { flatMap { listOf(it.π1, it.π3) }.toSet() }
+
+// FSAs looks like this:
+/*
+INIT -> 1 | 3
+DONE -> 4
+1 -<a>-> 1
+1 -<+>-> 3
+3 -<b>-> 4
+4 -<+>-> 1
+4 -<b>-> 4
+ */
+
+fun Σᐩ.parseFSA(): FSA {
+  val Q =
+    lines().asSequence()
+      .filter { it.isNotBlank() }
+      .map { it.split("->") }
+      .map { (lhs, rhs) ->
+        val src = lhs.tokenizeByWhitespace().first()
+        val dst = rhs.split("|").map { it.trim() }.toSet()
+        val sym = if ("-<" in lhs && lhs.endsWith(">"))
+          lhs.split("-<").last().dropLast(1) else ""
+
+        setOf(src) * setOf(sym) * dst
+      }.flatten().toList()
+      .onEach { println(it) }
+  val init = Q.filter { it.π1 == "INIT" }.map { it.π3 }.toSet()
+  val final = Q.filter { it.π1 == "DONE" }.map { it.π3 }.toSet()
+  return FSA(Q.filter { it.π1 !in setOf("INIT", "DONE") }.toSet(), init, final)
+}
diff --git a/src/commonMain/kotlin/ai/hypergraph/kaliningraph/parsing/SeqValiant.kt b/src/commonMain/kotlin/ai/hypergraph/kaliningraph/parsing/SeqValiant.kt
@@ -32,7 +32,7 @@ class PTree(val root: String = "ε", val branches: List<Π2A<PTree>> = listOf())
   // Samples instantaneously from the parse forest, but may return duplicates
   // and only returns a fraction of the number of distinct strings when compared
   // to SWOR on medium-sized finite sets under the same wall-clock timeout. If
-  // the set is sufficiently large, distinctness will never a problem.
+  // the set is sufficiently large, distinctness will never be a problem.
   fun sampleWithReplacement(): Sequence<String> = sequence { while(true) yield(sample()) }
 
   fun sample(): String =

diff --git a/src/commonTest/kotlin/ai/hypergraph/kaliningraph/parsing/BarHillelTest.kt b/src/commonTest/kotlin/ai/hypergraph/kaliningraph/parsing/BarHillelTest.kt
@@ -13,139 +13,6 @@ class BarHillelTest {
 */
 @Test
   fun testManualBarHillel() {
-    // Generated from https://github.com/breandan/bar-hillel/blob/7527d2ad1a007fb4667b85cec295a43c56a237db/rayuela/test/cfg/test_epsilon_Bar_Hillel.py
-    val bhcfg = """
-        START -> [1,START,4]
-        START -> [3,START,4]
-        [1,+,3] -> +
-        [1,+,3] -> *
-        [1,L,1] -> [1,O,1] [1,N,1]
-        [1,L,1] -> [1,O,3] [3,N,1]
-        [1,L,1] -> [1,O,4] [4,N,1]
-        [1,L,3] -> [1,O,1] [1,N,3]
-        [1,L,3] -> [1,O,3] [3,N,3]
-        [1,L,3] -> [1,O,4] [4,N,3]
-        [1,L,4] -> [1,O,1] [1,N,4]
-        [1,L,4] -> [1,O,3] [3,N,4]
-        [1,L,4] -> [1,O,4] [4,N,4]
-        [1,N,1] -> [1,a,1]
-        [1,N,1] -> [1,b,1]
-        [1,N,1] -> [1,N,1] [1,N,1]
-        [1,N,1] -> [1,N,3] [3,N,1]
-        [1,N,1] -> [1,N,4] [4,N,1]
-        [1,N,3] -> [1,a,3]
-        [1,N,3] -> [1,b,3]
-        [1,N,3] -> [1,N,1] [1,N,3]
-        [1,N,3] -> [1,N,3] [3,N,3]
-        [1,N,3] -> [1,N,4] [4,N,3]
-        [1,N,4] -> [1,a,4]
-        [1,N,4] -> [1,b,4]
-        [1,N,4] -> [1,N,1] [1,N,4]
-        [1,N,4] -> [1,N,3] [3,N,4]
-        [1,N,4] -> [1,N,4] [4,N,4]
-        [1,O,1] -> [1,x,1]
-        [1,O,1] -> [1,+,1]
-        [1,O,3] -> [1,x,3]
-        [1,O,3] -> [1,+,3]
-        [1,O,4] -> [1,x,4]
-        [1,O,4] -> [1,+,4]
-        [1,START,1] -> [1,N,1] [1,L,1]
-        [1,START,1] -> [1,N,3] [3,L,1]
-        [1,START,1] -> [1,N,4] [4,L,1]
-        [1,START,3] -> [1,N,1] [1,L,3]
-        [1,START,3] -> [1,N,3] [3,L,3]
-        [1,START,3] -> [1,N,4] [4,L,3]
-        [1,START,4] -> [1,N,1] [1,L,4]
-        [1,START,4] -> [1,N,3] [3,L,4]
-        [1,START,4] -> [1,N,4] [4,L,4]
-        [1,a,1] -> a
-        [3,L,1] -> [3,O,1] [1,N,1]
-        [3,L,1] -> [3,O,3] [3,N,1]
-        [3,L,1] -> [3,O,4] [4,N,1]
-        [3,L,3] -> [3,O,1] [1,N,3]
-        [3,L,3] -> [3,O,3] [3,N,3]
-        [3,L,3] -> [3,O,4] [4,N,3]
-        [3,L,4] -> [3,O,1] [1,N,4]
-        [3,L,4] -> [3,O,3] [3,N,4]
-        [3,L,4] -> [3,O,4] [4,N,4]
-        [3,N,1] -> [3,a,1]
-        [3,N,1] -> [3,b,1]
-        [3,N,1] -> [3,N,1] [1,N,1]
-        [3,N,1] -> [3,N,3] [3,N,1]
-        [3,N,1] -> [3,N,4] [4,N,1]
-        [3,N,3] -> [3,a,3]
-        [3,N,3] -> [3,b,3]
-        [3,N,3] -> [3,N,1] [1,N,3]
-        [3,N,3] -> [3,N,3] [3,N,3]
-        [3,N,3] -> [3,N,4] [4,N,3]
-        [3,N,4] -> [3,a,4]
-        [3,N,4] -> [3,b,4]
-        [3,N,4] -> [3,N,1] [1,N,4]
-        [3,N,4] -> [3,N,3] [3,N,4]
-        [3,N,4] -> [3,N,4] [4,N,4]
-        [3,O,1] -> [3,x,1]
-        [3,O,1] -> [3,+,1]
-        [3,O,3] -> [3,x,3]
-        [3,O,3] -> [3,+,3]
-        [3,O,4] -> [3,x,4]
-        [3,O,4] -> [3,+,4]
-        [3,START,1] -> [3,N,1] [1,L,1]
-        [3,START,1] -> [3,N,3] [3,L,1]
-        [3,START,1] -> [3,N,4] [4,L,1]
-        [3,START,3] -> [3,N,1] [1,L,3]
-        [3,START,3] -> [3,N,3] [3,L,3]
-        [3,START,3] -> [3,N,4] [4,L,3]
-        [3,START,4] -> [3,N,1] [1,L,4]
-        [3,START,4] -> [3,N,3] [3,L,4]
-        [3,START,4] -> [3,N,4] [4,L,4]
-        [3,b,4] -> b
-        [4,+,1] -> +
-        [4,+,1] -> *
-        [4,L,1] -> [4,O,1] [1,N,1]
-        [4,L,1] -> [4,O,3] [3,N,1]
-        [4,L,1] -> [4,O,4] [4,N,1]
-        [4,L,3] -> [4,O,1] [1,N,3]
-        [4,L,3] -> [4,O,3] [3,N,3]
-        [4,L,3] -> [4,O,4] [4,N,3]
-        [4,L,4] -> [4,O,1] [1,N,4]
-        [4,L,4] -> [4,O,3] [3,N,4]
-        [4,L,4] -> [4,O,4] [4,N,4]
-        [4,N,1] -> [4,a,1]
-        [4,N,1] -> [4,b,1]
-        [4,N,1] -> [4,N,1] [1,N,1]
-        [4,N,1] -> [4,N,3] [3,N,1]
-        [4,N,1] -> [4,N,4] [4,N,1]
-        [4,N,3] -> [4,a,3]
-        [4,N,3] -> [4,b,3]
-        [4,N,3] -> [4,N,1] [1,N,3]
-        [4,N,3] -> [4,N,3] [3,N,3]
-        [4,N,3] -> [4,N,4] [4,N,3]
-        [4,N,4] -> [4,a,4]
-        [4,N,4] -> [4,b,4]
-        [4,N,4] -> [4,N,1] [1,N,4]
-        [4,N,4] -> [4,N,3] [3,N,4]
-        [4,N,4] -> [4,N,4] [4,N,4]
-        [4,O,1] -> [4,x,1]
-        [4,O,1] -> [4,+,1]
-        [4,O,3] -> [4,x,3]
-        [4,O,3] -> [4,+,3]
-        [4,O,4] -> [4,x,4]
-        [4,O,4] -> [4,+,4]
-        [4,START,1] -> [4,N,1] [1,L,1]
-        [4,START,1] -> [4,N,3] [3,L,1]
-        [4,START,1] -> [4,N,4] [4,L,1]
-        [4,START,3] -> [4,N,1] [1,L,3]
-        [4,START,3] -> [4,N,3] [3,L,3]
-        [4,START,3] -> [4,N,4] [4,L,3]
-        [4,START,4] -> [4,N,1] [1,L,4]
-        [4,START,4] -> [4,N,3] [3,L,4]
-        [4,START,4] -> [4,N,4] [4,L,4]
-        [4,b,4] -> b
-      """.trimIndent().parseCFG().noNonterminalStubs
-
-    println(bhcfg.pretty)
-
-    // bhcfg is the intersection of
     val cfg = """
         START -> N L
         N -> N N | a | b
@@ -155,13 +22,18 @@ class BarHillelTest {
       """.parseCFG().noNonterminalStubs
 
     val fsa = """
-        INIT: 1, 3 FINAL: 4
-        1 -[a]-> 1
-        1 -[+]-> 3
-        3 -[b]-> 4
-        4 -[+]-> 1
-        4 -[b]-> 4
-      """.trimIndent()
+        INIT -> 1 | 3
+        DONE -> 4
+        1 -<a>-> 1
+        1 -<+>-> 3
+        1 -<*>-> 3
+        3 -<b>-> 4
+        4 -<+>-> 1
+        4 -<*>-> 1
+        4 -<b>-> 4
+      """.parseFSA()
+
+    val bhcfg = cfg.intersect(fsa)
 
     val fsaCfg = """
         START -> START b | 3 b
@@ -191,7 +63,7 @@ class BarHillelTest {
         assertTrue { it in bhcfg.language }
         assertTrue { it in fsaCfg.language }
         assertTrue { it in cfg.language }
-      }.take(300).toList().also { println("Found ${it.size} solutions.") }
+      }.take(300).toList().also { println("Sampling solver found ${it.size} solutions.") }
     }.also { println("Sampling solver took: ${it.inWholeMilliseconds}ms") }
 
     clock = TimeSource.Monotonic.markNow()
@@ -201,7 +73,7 @@ class BarHillelTest {
         assertTrue { it in bhcfg.language }
         assertTrue { it in fsaCfg.language }
         assertTrue { it in cfg.language }
-      }.toList().also { println("Found ${it.size} solutions.") }
+      }.toList().also { println("Sequential solver found ${it.size} solutions.") }
     }.also { println("Sequential solver took: ${it.inWholeMilliseconds}ms") }
 
     clock = TimeSource.Monotonic.markNow()
@@ -211,7 +83,7 @@ class BarHillelTest {
         assertTrue { it in bhcfg.language }
         assertTrue { it in fsaCfg.language }
         assertTrue { it in cfg.language }
-      }.toList().also { println("Found ${it.size} solutions.") }
+      }.toList().also { println("Sort solver found ${it.size} solutions.") }
     }.also { println("Sort solver took: ${it.inWholeMilliseconds}ms") }
   }