powerpoint

Create a Presentation object, add slides with prs.slide_layouts[6] for blank layouts, then add shapes and textboxes with Inches-based coordinates. Write the script to a file and run it with system Python (/usr/bin/python3), since sandbox venvs often lack python-pptx.

Modify the sldIdLst element inside ppt/presentation.xml directly via zipfile, because python-pptx maintains two internal references and only the part-level XML is saved. The included scripts/reorder_pptx.py automates this with a comma-separated index order.

python-pptx suits Python-centric workflows with direct shape and table control, while pptxgenjs fits Node.js environments with built-in charts and slide masters. Both are documented here with their specific pitfalls, such as hex color corruption rules in pptxgenjs.

LibreOffice clips elements beyond the 12.8 x 7.0 inch printable area even though slides are 13.333 x 7.5 inches. Multiply all non-zero Inches() coordinates by 0.92 and verify with scripts/quality_check.py before delivery.

The _Cell object has no paragraphs attribute; you must access cell.text_frame.paragraphs instead. Also set cell.text_frame.word_wrap = True explicitly, since wrapping is not enabled by default.

Read the original dimensions with PIL, compute the aspect ratio, fit the image within the target area by width or height, and center it before calling add_picture. Passing width and height directly stretches the image.