Huayu Hanako Academic PPT Generator

Problem

Turning a long academic manuscript into slides is not mainly a copy-paste task. The harder part is deciding which sentences form the central argument, which tables should become charts, which sections should be merged, and where restraint improves clarity. Template-driven decks often compress reasoning into bullet points or add decoration without improving logical flow.

Workflow

The skill first reads the input text and extracts the core argument, structure, key data, section weight, and logical links. It then produces a 25-30 page outline and a runnable Python script that writes a .pptx file via python-pptx. The script emphasizes readable annotations and error handling for local reproduction and modification. It supports six themes, including Morandi blue, green, clay, mauve, gray, and a classic business green, with consistent top bars, title panels, quote blocks, code blocks, and table styling. When tabular data is detected, it can generate pie, bar, or line charts; images are fetched with a four-level fallback across Pexels, Baidu, Unsplash, and Pixabay.

Boundaries

Output depends on a stable network and a Python 3.8+ environment. Image APIs have rate limits, keyword matching may not perfectly suit every academic context, and the table of contents is currently static text without automatic slide navigation.