Skip to content

Fix TextChunker orphan chunk merging to use token count instead of word count #13714

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
joaquinhuigomez wants to merge 2 commits into
base: main
Choose a base branch
from
+6 −11
Open

Fix TextChunker orphan chunk merging to use token count instead of word count #13714

Changes from 1 commit
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
08a3780
Fix TextChunker orphan chunk gluing using word count instead of token…
joaquinhuigomez Mar 27, 2026
4a2b7b5
Hoist lastParagraphTokenCount to avoid redundant token counter call
joaquinhuigomez Mar 27, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
14 changes: 4 additions & 10 deletions dotnet/src/SemanticKernel.Core/Text/TextChunker.cs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -192,18 +192,12 @@ private static List<string> ProcessParagraphs(List<string> paragraphs, int adjus

if (GetTokenCount(lastParagraph, tokenCounter) < adjustedMaxTokensPerParagraph / 4)
{
var lastParagraphTokens = lastParagraph.Split(s_spaceChar, StringSplitOptions.RemoveEmptyEntries);
var secondLastParagraphTokens = secondLastParagraph.Split(s_spaceChar, StringSplitOptions.RemoveEmptyEntries);
var lastParagraphTokenCount = GetTokenCount(lastParagraph, tokenCounter);
Copy link
Copy Markdown
Contributor

@github-actions github-actions bot Mar 27, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Nit: GetTokenCount(lastParagraph, tokenCounter) is already computed on line 193 for the if guard, then recomputed identically here. Hoist the result into a local before the if to avoid invoking the token counter twice — especially relevant when tokenCounter is a non-trivial delegate.

Suggested change
if (GetTokenCount(lastParagraph, tokenCounter) < adjustedMaxTokensPerParagraph / 4)
{
var lastParagraphTokens = lastParagraph.Split(s_spaceChar, StringSplitOptions.RemoveEmptyEntries);
var secondLastParagraphTokens = secondLastParagraph.Split(s_spaceChar, StringSplitOptions.RemoveEmptyEntries);
var lastParagraphTokenCount = GetTokenCount(lastParagraph, tokenCounter);
var lastParagraphTokenCount = GetTokenCount(lastParagraph, tokenCounter);
if (lastParagraphTokenCount < adjustedMaxTokensPerParagraph / 4)
{

var secondLastParagraphTokenCount = GetTokenCount(secondLastParagraph, tokenCounter);

var lastParagraphTokensCount = lastParagraphTokens.Length;
var secondLastParagraphTokensCount = secondLastParagraphTokens.Length;

if (lastParagraphTokensCount + secondLastParagraphTokensCount <= adjustedMaxTokensPerParagraph)
if (lastParagraphTokenCount + secondLastParagraphTokenCount <= adjustedMaxTokensPerParagraph)
{
var newSecondLastParagraph = string.Join(" ", secondLastParagraphTokens);
var newLastParagraph = string.Join(" ", lastParagraphTokens);

paragraphs[paragraphs.Count - 2] = $"{newSecondLastParagraph} {newLastParagraph}";
paragraphs[paragraphs.Count - 2] = $"{secondLastParagraph} {lastParagraph}";
paragraphs.RemoveAt(paragraphs.Count - 1);
}
}
Expand Down
Loading