Keep the data points of XY and bubble chart tables

packages/markitdown/src/markitdown/converters/_pptx_converter.py

@@ -343,6 +343,14 @@ def _convert_chart_to_markdown(self, chart):
             # XPath scan of all points), which is extremely slow on large charts.
             series_values = [list(s.values) for s in series_list]
 
+            # XY (scatter) and bubble charts have no category axis, so their
+            # data points only exist in the series themselves. Reuse the point
+            # position as the row label; otherwise every value below would be
+            # dropped along with the missing categories.
+            if not category_names:
+                point_count = max((len(sv) for sv in series_values), default=0)
+                category_names = [str(idx + 1) for idx in range(point_count)]
+
             for idx, category in enumerate(category_names):
                 row = [category]
                 for sv in series_values:

packages/markitdown/tests/test_pptx_chart_no_categories.py

@@ -0,0 +1,74 @@
+"""PowerPoint charts that have no category axis (XY scatter and bubble).
+
+Those charts keep their data points inside the series as ``(x, y)`` pairs, so
+``chart.plots[0].categories`` is empty and the converter used to emit a table
+that ended right after its separator line, dropping every value.
+"""
+
+import io
+
+import pptx
+from pptx.chart.data import BubbleChartData, XyChartData
+from pptx.enum.chart import XL_CHART_TYPE
+from pptx.util import Inches
+
+from markitdown import MarkItDown, StreamInfo
+
+SCATTER_VALUES = [10.0, 20.0, 30.0]
+
+
+def _chart_presentation(chart_type, chart_data) -> io.BytesIO:
+    """Build a one-slide deck holding a single chart titled "Sales"."""
+    prs = pptx.Presentation()
+    slide = prs.slides.add_slide(prs.slide_layouts[6])
+    chart = slide.shapes.add_chart(
+        chart_type, Inches(1), Inches(1), Inches(6), Inches(4), chart_data
+    ).chart
+    chart.has_title = True
+    chart.chart_title.text_frame.text = "Sales"
+    buf = io.BytesIO()
+    prs.save(buf)
+    buf.seek(0)
+    return buf
+
+
+def _scatter_presentation() -> io.BytesIO:
+    """Build a one-slide deck holding an XY (scatter) chart."""
+    chart_data = XyChartData()
+    series = chart_data.add_series("Revenue")
+    for idx, value in enumerate(SCATTER_VALUES, start=1):
+        series.add_data_point(float(idx), value)
+    return _chart_presentation(XL_CHART_TYPE.XY_SCATTER, chart_data)
+
+
+def _bubble_presentation() -> io.BytesIO:
+    """Build a one-slide deck holding a bubble chart."""
+    chart_data = BubbleChartData()
+    series = chart_data.add_series("Revenue")
+    for idx, value in enumerate(SCATTER_VALUES, start=1):
+        series.add_data_point(float(idx), value, 1.0)
+    return _chart_presentation(XL_CHART_TYPE.BUBBLE, chart_data)
+
+
+def _convert(stream: io.BytesIO) -> str:
+    return (
+        MarkItDown()
+        .convert_stream(stream, stream_info=StreamInfo(extension=".pptx"))
+        .markdown
+    )
+
+
+def test_pptx_scatter_chart_keeps_data_points() -> None:
+    markdown = _convert(_scatter_presentation())
+
+    assert "### Chart: Sales" in markdown
+    for idx, value in enumerate(SCATTER_VALUES, start=1):
+        assert f"| {idx} | {value} |" in markdown
+
+
+def test_pptx_bubble_chart_keeps_data_points() -> None:
+    markdown = _convert(_bubble_presentation())
+
+    assert "### Chart: Sales" in markdown
+    for idx, value in enumerate(SCATTER_VALUES, start=1):
+        assert f"| {idx} | {value} |" in markdown