fix(utils): recurse into PPTX groups and tables when extracting text (#3250)

nanobot/utils/document.py

@@ -159,8 +159,7 @@ def _extract_pptx(path: Path) -> str:
         for i, slide in enumerate(prs.slides, 1):
             slide_text: list[str] = []
             for shape in slide.shapes:
-                if hasattr(shape, "text") and shape.text:
-                    slide_text.append(shape.text)
+                _collect_pptx_shape_text(shape, slide_text)
             if slide_text:
                 slides.append(f"--- Slide {i} ---\n" + "\n".join(slide_text))
         return _truncate("\n\n".join(slides), _MAX_TEXT_LENGTH)
@@ -169,6 +168,31 @@ def _extract_pptx(path: Path) -> str:
         return f"[error: failed to extract PPTX: {e!s}]"
 
 
+def _collect_pptx_shape_text(shape, out: list[str]) -> None:
+    """Collect text from a PPTX shape, recursing into groups and tables.
+
+    Groups have ``has_text_frame=False`` and must be walked via ``.shapes``;
+    tables are GraphicFrame objects whose cell text lives under ``.table``.
+    """
+    sub_shapes = getattr(shape, "shapes", None)
+    if sub_shapes is not None:
+        for sub in sub_shapes:
+            _collect_pptx_shape_text(sub, out)
+        return
+
+    if getattr(shape, "has_table", False):
+        for row in shape.table.rows:
+            cells = [cell.text.strip() for cell in row.cells]
+            line = "\t".join(cell for cell in cells if cell)
+            if line:
+                out.append(line)
+        return
+
+    text = getattr(shape, "text", "")
+    if text:
+        out.append(text)
+
+
 def _extract_text_file(path: Path) -> str:
     """Extract text from a plain text file."""
     try:

tests/test_document_parsing.py

@@ -209,6 +209,49 @@ class TestExtractText:
         # Text content may vary depending on PowerPoint layout defaults
         assert len(result) > 0
 
+    def test_extract_text_pptx_table(self, tmp_path: Path):
+        """Table cells should be extracted, not silently dropped."""
+        from pptx import Presentation
+        from pptx.util import Inches
+
+        pptx_file = tmp_path / "table.pptx"
+        prs = Presentation()
+        slide = prs.slides.add_slide(prs.slide_layouts[5])
+        table = slide.shapes.add_table(
+            2, 2, Inches(1), Inches(1), Inches(4), Inches(1)
+        ).table
+        table.cell(0, 0).text = "Header A"
+        table.cell(0, 1).text = "Header B"
+        table.cell(1, 0).text = "Alice"
+        table.cell(1, 1).text = "Bob"
+        prs.save(pptx_file)
+
+        result = extract_text(pptx_file)
+        assert result is not None
+        assert "Header A" in result
+        assert "Header B" in result
+        assert "Alice" in result
+        assert "Bob" in result
+
+    def test_extract_text_pptx_grouped_shapes(self, tmp_path: Path):
+        """Text inside grouped shapes must be extracted recursively."""
+        from pptx import Presentation
+        from pptx.util import Inches
+
+        pptx_file = tmp_path / "grouped.pptx"
+        prs = Presentation()
+        slide = prs.slides.add_slide(prs.slide_layouts[5])
+        group = slide.shapes.add_group_shape()
+        inner = group.shapes.add_textbox(
+            Inches(1), Inches(1), Inches(3), Inches(1)
+        )
+        inner.text_frame.text = "Inside group"
+        prs.save(pptx_file)
+
+        result = extract_text(pptx_file)
+        assert result is not None
+        assert "Inside group" in result
+
     def test_extract_text_pdf_not_found(self, tmp_path: Path):
         """Test that missing PDF files return error string."""
         missing_pdf = tmp_path / "nonexistent.pdf"