fix(chunkers): address PR review comments

- Fix regex fallback path: use sliding window for overlap instead of
  passing chunkOverlap to buildChunks without prepended overlap text
- Fix misleading strategy label: "Text (hierarchical splitting)" →
  "Text (word boundary splitting)"

Author: waleedlatif1

apps/sim/app/workspace/[workspaceId]/knowledge/components/create-base-modal/create-base-modal.tsx

@@ -40,7 +40,7 @@ interface CreateBaseModalProps {
 
 const STRATEGY_OPTIONS = [
   { value: 'auto', label: 'Auto (detect from content)' },
-  { value: 'text', label: 'Text (hierarchical splitting)' },
+  { value: 'text', label: 'Text (word boundary splitting)' },
   { value: 'recursive', label: 'Recursive (configurable separators)' },
   { value: 'sentence', label: 'Sentence' },
   { value: 'token', label: 'Token (fixed-size)' },

apps/sim/lib/chunkers/regex-chunker.ts

@@ -86,8 +86,10 @@ export class RegexChunker {
     if (segments.length <= 1) {
       logger.warn('Regex pattern did not produce any splits, falling back to character splitting')
       const chunkSizeChars = tokensToChars(this.chunkSize)
-      const chunks = splitAtWordBoundaries(cleaned, chunkSizeChars)
-      return buildChunks(chunks, this.chunkOverlap)
+      const overlapChars = tokensToChars(this.chunkOverlap)
+      const stepChars = this.chunkOverlap > 0 ? chunkSizeChars - overlapChars : undefined
+      const chunks = splitAtWordBoundaries(cleaned, chunkSizeChars, stepChars)
+      return buildChunks(chunks, 0)
     }
 
     const merged = this.mergeSegments(segments)