mumford

Format-aware diff engines for common document types — PDF, Word, Excel, RTF, PowerPoint, JSON, plain text. Built on top of tate.

Overview

Mumford provides file-format parsing and diffing for real-world documents. All diff algorithms come from tate; mumford handles the format-specific work of reading files and converting them into the data structures tate operates on.

Supported formats

Format	Extensions	Engine
Plain text	.txt, .csv, .log, .md, code files	Line-level diff via tate::lines
PDF	.pdf	Text extraction via pdfium, running-header stripping, line diff
Word	.docx	OOXML paragraph + table extraction, paragraph diff
Excel	.xlsx, .xls, .xlsm	Cell-level grid alignment via tate::grid
RTF	.rtf	Styled-table parsing, row/cell diff with formatting preserved
PowerPoint	.pptx	Slide text extraction, slide-level alignment + line diff
JSON	.json	Structural diff via tate::tree::TreeNode
Folders	directories	Recursive comparison with sha256 hashing and rename detection

Architecture

tate (algorithms: lines, inline, grid, tree)
  ↑
mumford (format engines: parse → feed to tate → wrap result)
  ↑
your app (domain adapters + UI)

Usage

[dependencies]
mumford = "0.1"

Diff two files (auto-dispatch by extension)

use mumford::dispatch;

let result = dispatch("old.xlsx", "new.xlsx")?;
if let Some(excel) = result.excel {
    for sheet in &excel.sheets {
        println!("{}: {} rows modified", sheet.name, sheet.grid.modified_rows);
    }
}

JSON structural diff

use mumford::json::json_diff;

let diff = json_diff(r#"{"port": 8080}"#, r#"{"port": 9090}"#)?;
assert_eq!(diff.changes[0].id, "port");

Notes

• PDF requires pdfium: The native pdfium library must be available at runtime (via PDFIUM_LIB_PATH, next to the executable, or the system loader). See bblanchon/pdfium-binaries for pre-built binaries.

License

MIT