[Limdongcheng]

Summary

XLSX import, in the desktop and through data.import, reads journal Source Data worksheets as clean tables:

• Splits side-by-side panels at blank columns and at first-row panel letters, naming tables like Figure 3 b.
• Joins multi-row headers into column names, reads single-cell rows as titles and shares sample names across each repeating x/y header group.
• Reads placeholders such as --, NaN, N/A and n.d., and other text among values, as missing in numeric columns, and ignores notes below the values.
• Splits stacked segments, such as charge and discharge separated by a repeated header, into numbered columns aligned with their x column when every column of the group pauses on those rows.
• data.import accepts regions (sheet, A1 range, header_rows, name, split_segments) to import explicit ranges, and reports each table's range. A region or table that fails fails on its own.

Worksheets with one row of distinct text headers over one block of plain data import as before.

Validation

• cargo pr-check
• npm run build in docs/
• Imported the Source Data workbooks of ten published papers with plotium-cli tool run.