ZM Research Report

Problem

Research figures often sit in a folder as PDF/TIFF/PNG files. Manually assembling them into a report can cause missing figures, inconsistent figure captions, uneven subscripts/superscripts, and repeated thesis-style formatting. This skill turns the folder-to-Chinese-.docx workflow into a repeatable script.

How It Works

Given a folder containing description.txt and figure files, it:

• Reads the input: scans the folder, prefers PDF/TIFF vector figures, and sorts them by filename; if only PNG rasters exist, it asks before continuing; if description.txt is missing, it keeps only images and section scaffolding with placeholders.
• Generates gen_report.py: uses python-docx to produce a .docx with a doctoral thesis layout: A4 page, SimSun small-4, Times New Roman, 1.5 line spacing, bold headings, “Chapter X” numbering, and centered “Figure X.Y” captions.
• Handles sub/superscripts: converts Unicode sub/superscripts and αk patterns, parses markers, splits runs, and sets font.subscript / font.superscript.
• Constrains writing style: uses academic Chinese, avoids “解耦” and “耦合”, pairs each figure with an independent analysis that reports r/p values before explaining mechanisms, and adds a group summary.
• Manages versions and checks: increments report versions, keeps old .docx files, and prints Images embedded: N, where N must match the total figure count.

Boundaries

It is suited for formatting existing figures and descriptions into a standardized Chinese report, not for designing experiments or interpreting conclusions. Without description.txt, the report is only a skeleton with placeholders; if certain figure sets are missing, the skill may try to fill gaps or prompt, but the input folder still needs the source figures.