try using markov chain to select candidate fragments

breandan · Sep 22, 2023 · df725b7 · df725b7
1 parent 83bc0a6
commit df725b7
Show file tree

Hide file tree

Showing 5 changed files with 18 additions and 13 deletions.
diff --git a/src/commonMain/kotlin/ai/hypergraph/kaliningraph/parsing/CFG.kt b/src/commonMain/kotlin/ai/hypergraph/kaliningraph/parsing/CFG.kt
@@ -5,6 +5,7 @@ import ai.hypergraph.kaliningraph.graphs.LabeledGraph
 import ai.hypergraph.kaliningraph.sampling.choose
 import ai.hypergraph.kaliningraph.types.*
 import kotlin.jvm.JvmName
+import kotlin.reflect.KProperty
 import kotlin.time.*
 
 typealias Σᐩ = String

diff --git a/src/commonMain/kotlin/ai/hypergraph/kaliningraph/parsing/SetValiantSpeedy.kt b/src/commonMain/kotlin/ai/hypergraph/kaliningraph/parsing/SetValiantSpeedy.kt
@@ -76,8 +76,7 @@ fun FastRandomSet<Edit>.resample(maxTake: Int,
 
 // Enumerates powerset levels from the bottom up, skipping the empty set
 private fun Edit.subedits(): Sequence<Sequence<List<Pair<Int, Σᐩ>>>> =
-  (1..size).asSequence()
-  .map { choose(it).map { it.toList() } }
+  (1..size).asSequence().map { choose(it).map { it.toList() } }
 
 fun List<Σᐩ>.apply(edit: Edit): List<Σᐩ> {
   val res = toMutableList()

diff --git a/src/commonMain/kotlin/ai/hypergraph/kaliningraph/parsing/SortValiant.kt b/src/commonMain/kotlin/ai/hypergraph/kaliningraph/parsing/SortValiant.kt
@@ -7,10 +7,12 @@ import ai.hypergraph.kaliningraph.tensor.UTMatrix
 import ai.hypergraph.kaliningraph.types.*
 
 // Returns all syntactically strings ordered by distance to withRespect
-fun CFG.sortAll(s: Σᐩ, metric: ChoiceMetric): Set<Σᐩ> =
-  try { solveSortedFP(s.tokenizeByWhitespace(), metric)?.sorted()
-    ?.map { it.asString }?.toSet() ?: setOf() }
-  catch (e: Exception) { e.printStackTrace(); setOf() }
+fun CFG.solve(s: Σᐩ, metric: ChoiceMetric): Set<Σᐩ> =
+  solve(s.tokenizeByWhitespace(), metric)
+
+fun CFG.solve(s: List<Σᐩ>, metric: ChoiceMetric): Set<Σᐩ> =
+  try { solveSortedFP(s, metric)?.sorted()?.map { it.asString }?.toSet() }
+  catch (e: Exception) { e.printStackTrace(); null } ?: setOf()
 
 fun CFG.solveSortedFP(
   tokens: List<Σᐩ>,
@@ -43,6 +45,7 @@ fun CFG.sortwiseAlgebra(metric: ChoiceMetric): Ring<Sort> =
 
 const val MAX_CAPACITY = 100
 // 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 =
   bimap.TRIPL.filter { (_, x, z) -> x in X && z in Z }
   .map { (w, x, z) ->

diff --git a/src/commonTest/kotlin/ai/hypergraph/kaliningraph/parsing/SetValiantTest.kt b/src/commonTest/kotlin/ai/hypergraph/kaliningraph/parsing/SetValiantTest.kt
@@ -1,6 +1,6 @@
 package ai.hypergraph.kaliningraph.parsing
 
-import ai.hypergraph.kaliningraph.hasBalancedBrackets
+import ai.hypergraph.kaliningraph.*
 import ai.hypergraph.kaliningraph.tensor.seekFixpoint
 import ai.hypergraph.kaliningraph.types.π2
 import kotlinx.datetime.Clock
@@ -363,7 +363,7 @@ class SetValiantTest {
     val holExpr = "_ _ _ _ _ _ _ _ _ _"
 
     measureTime {
-      val solutions = ocamlCFG.sortAll(holExpr, levMetric("( false curry )"))
+      val solutions = ocamlCFG.solve(holExpr, levMetric("( false curry )"))
       println("Found: ${solutions.size} unique solutions")
       solutions.forEach { println(it); assertTrue("$it was invalid!") { ocamlCFG.isValid(it) } }
     }.also { println("Finished in ${it.inWholeMilliseconds}ms.") }
@@ -569,11 +569,12 @@ Yield_Arg -> From_Keyword Test | Testlist_Endcomma
 */
   @Test
   fun testPythonRepairs() {
-    val reference = "NAME = ( NAME"
-    val template = "_ _ _ _ _ _ _"//List(it.size + 2) { "_" }.joinToString(" ")
+    val refStr = "NAME = ( NAME"
+    val refLst = refStr.tokenizeByWhitespace()
+    val template = List(refLst.size + 3) { "_" }.joinToString(" ")
     measureTime {
-      seq2parsePythonCFG.sortAll(template, levMetric(reference))
-        .onEach { println(it) }
+      seq2parsePythonCFG.solve(template, levMetric(refStr))
+        .onEach { println("Δ=${levenshtein(it, refStr)}: $it") }
         .also { println("Found ${it.size} solutions!") }
     }.also { println("Finished in ${it.inWholeMilliseconds}ms.") }
   }

diff --git a/src/jvmMain/kotlin/ai/hypergraph/markovian/mcmc/MarkovChain.kt b/src/jvmMain/kotlin/ai/hypergraph/markovian/mcmc/MarkovChain.kt
@@ -114,9 +114,10 @@ open class MarkovChain<T>(
 
   // Computes perplexity of a sequence normalized by sequence length
   fun score(seq: List<T>): Double =
-    -seq.windowed(memory)
+    if (memory < seq.size) -seq.windowed(memory)
       .map { (getAtLeastOne(it) + 1) / (getAtLeastOne(it.dropLast(1) + null) + dictionary.size) }
       .sumOf { ln(it) } / seq.size
+    else (seq.sumOf { counter.rawCounts.getEstimate(it) } + 1).toDouble() / counter.total.toDouble()
 
   operator fun get(vararg variables: T?): Double =
     if (variables.size == 1) counter.rawCounts.getEstimate(variables[0]) / counter.total.toDouble()