From 28bc10df929f603e671d2cefeabb196fbec2caa6 Mon Sep 17 00:00:00 2001
From: Daniel O'Connell <github@ahiru.pl>
Date: Fri, 19 Dec 2025 18:27:13 +0100
Subject: [PATCH] Fix break_chunk() appending wrong object (BUG-007)
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

The function was appending the entire DataChunk object instead of
the individual item when processing non-string data (e.g., images).

Bug: `result.append(chunk)` should have been `result.append(c)`

This caused:
- Type mismatches (returning DataChunk instead of MulitmodalChunk)
- Potential circular references
- Embedding failures for mixed content

Fixed by appending the individual item `c` instead of the parent `chunk`.
Updated existing test and added new test to verify behavior.

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
---
 src/memory/common/embedding.py        |  5 +++--
 tests/memory/common/test_embedding.py | 24 +++++++++++++++++++++---
 2 files changed, 24 insertions(+), 5 deletions(-)

diff --git a/src/memory/common/embedding.py b/src/memory/common/embedding.py
index 9a4e752..3fc1f20 100644
--- a/src/memory/common/embedding.py
+++ b/src/memory/common/embedding.py
@@ -49,12 +49,13 @@ def embed_chunks(
 def break_chunk(
     chunk: extract.DataChunk, chunk_size: int = DEFAULT_CHUNK_TOKENS
 ) -> list[extract.MulitmodalChunk]:
-    result = []
+    result: list[extract.MulitmodalChunk] = []
     for c in chunk.data:
         if isinstance(c, str):
             result += chunk_text(c, chunk_size, OVERLAP_TOKENS)
         else:
-            result.append(chunk)
+            # Non-string items (e.g., images) are passed through directly
+            result.append(c)
     return result
 
 
diff --git a/tests/memory/common/test_embedding.py b/tests/memory/common/test_embedding.py
index 7e45a59..3407dce 100644
--- a/tests/memory/common/test_embedding.py
+++ b/tests/memory/common/test_embedding.py
@@ -153,11 +153,29 @@ def test_break_chunk_with_mixed_data_types():
     chunk = DataChunk(data=["text content", mock_image])
     result = break_chunk(chunk, chunk_size=100)
 
-    # Should have text chunks plus the original chunk (since it's not a string)
+    # Should have text chunks plus the image (non-string items are passed through)
     assert len(result) >= 2
     assert any(isinstance(item, str) for item in result)
-    # The original chunk should be preserved when it contains mixed data
-    assert chunk in result
+    # The individual non-string item (image) should be in result, not the DataChunk
+    assert mock_image in result
+    # The DataChunk itself should NOT be in the result
+    assert chunk not in result
+
+
+def test_break_chunk_preserves_non_string_items():
+    """Non-string items (like images) should be preserved individually."""
+    mock_image1 = Mock(spec=Image.Image)
+    mock_image2 = Mock(spec=Image.Image)
+    chunk = DataChunk(data=[mock_image1, "some text", mock_image2])
+    result = break_chunk(chunk, chunk_size=100)
+
+    # Both images should be in result
+    assert mock_image1 in result
+    assert mock_image2 in result
+    # Text should be chunked
+    assert "some text" in result
+    # Total should be 3 items (2 images + 1 short text)
+    assert len(result) == 3
 
 
 def test_embed_by_model_with_matching_chunks(mock_embed):