blun-king-cli 9.1.204 → 9.1.205

package/bin/repeated-assistant-response-policy.cjs

@@ -8,6 +8,7 @@ const REPEATED_RESPONSE_PLACEHOLDER = '[Historical repeated assistant response o
 const INTRA_MESSAGE_MIN_REPEATS = 3;
 const INTRA_MESSAGE_MIN_BLOCK_CHARS = 40;
 const INTRA_MESSAGE_MIN_SAVED_CHARS = 1_000;
+const INTRA_MESSAGE_SIMILARITY_THRESHOLD = 0.88;
 const INTRA_MESSAGE_REPEAT_PLACEHOLDER = '[Exact repeated assistant paragraphs omitted from model projection; raw history preserved.]';
 
 function isUserPrompt(message) {
@@ -60,6 +61,17 @@ function responseSimilarity(left, right) {
   return (2 * intersection) / (leftBigrams.size + rightBigrams.size);
 }
 
+function fencedBlockIndexes(blocks) {
+  const protectedIndexes = new Set();
+  let insideFence = false;
+  for (let index = 0; index < blocks.length; index += 1) {
+    const fenceCount = blocks[index].match(/(?:^|\n)\s*(?:```|~~~)/gu)?.length ?? 0;
+    if (insideFence || fenceCount > 0) protectedIndexes.add(index);
+    if (fenceCount % 2 === 1) insideFence = !insideFence;
+  }
+  return protectedIndexes;
+}
+
 function hasUserPromptBetween(history, leftIndex, rightIndex) {
   for (let index = leftIndex + 1; index < rightIndex; index += 1) {
     if (isUserPrompt(history[index])) return true;
@@ -120,27 +132,31 @@ function compactRepeatedAssistantParagraphs(message) {
   )) return null;
 
   const text = textParts[0].text;
-  if (/```|~~~/u.test(text)) return null;
   const blocks = text.split(/\n\s*\n/u).map((block) => block.trim()).filter(Boolean);
   if (blocks.length < INTRA_MESSAGE_MIN_REPEATS + 1) return null;
+  const protectedIndexes = fencedBlockIndexes(blocks);
+
+  const clusters = [];
+  for (let index = 0; index < blocks.length; index += 1) {
+    const block = blocks[index];
+    if (protectedIndexes.has(index) || block.length < INTRA_MESSAGE_MIN_BLOCK_CHARS) continue;
+    const cluster = clusters.find((candidate) => (
+      candidate.representative === block
+      || responseSimilarity(candidate.representative, block) >= INTRA_MESSAGE_SIMILARITY_THRESHOLD
+    ));
+    if (cluster) cluster.indexes.push(index);
+    else clusters.push({ indexes: [index], representative: block });
+  }
+  const repeatedIndexes = new Set(clusters
+    .filter((cluster) => cluster.indexes.length >= INTRA_MESSAGE_MIN_REPEATS)
+    .flatMap((cluster) => cluster.indexes.slice(1)));
+  if (repeatedIndexes.size === 0) return null;
 
-  const counts = new Map();
-  for (const block of blocks) counts.set(block, (counts.get(block) ?? 0) + 1);
-  const repeated = new Set([...counts]
-    .filter(([block, count]) => (
-      block.length >= INTRA_MESSAGE_MIN_BLOCK_CHARS
-      && count >= INTRA_MESSAGE_MIN_REPEATS
-    ))
-    .map(([block]) => block));
-  if (repeated.size === 0) return null;
-
-  const seen = new Set();
   const projectedBlocks = [];
   let markerAdded = false;
-  for (const block of blocks) {
-    if (!repeated.has(block) || !seen.has(block)) {
-      projectedBlocks.push(block);
-      if (repeated.has(block)) seen.add(block);
+  for (let index = 0; index < blocks.length; index += 1) {
+    if (!repeatedIndexes.has(index)) {
+      projectedBlocks.push(blocks[index]);
       continue;
     }
     if (!markerAdded) {
@@ -184,6 +200,7 @@ module.exports = {
   INTRA_MESSAGE_MIN_BLOCK_CHARS,
   INTRA_MESSAGE_MIN_REPEATS,
   INTRA_MESSAGE_MIN_SAVED_CHARS,
+  INTRA_MESSAGE_SIMILARITY_THRESHOLD,
   INTRA_MESSAGE_REPEAT_PLACEHOLDER,
   MAX_RECENT_ASSISTANT_RESPONSES,
   MIN_RESPONSE_CHARS,

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "blun-king-cli",
-  "version": "9.1.204",
+  "version": "9.1.205",
   "description": "BLUN CLI - your own AI agent with a Telegram channel. Get it done. With BLUN.",
   "license": "MIT",
   "bin": {