Research Report Kit

A standalone, local-first CSV quality inspector for people who need to understand a table before cleaning, reporting, or analysing it.

It is part of a local productivity-tool portfolio, but it remains independent from the Desktop File Organizer: separate repository, runtime, tests, output and release lifecycle.

What it checks

• encoding and delimiter
• row and column counts
• missing values by column
• duplicate rows
• irregular row widths
• duplicate or empty header labels
• numeric versus text columns
• numeric minimum, maximum, mean and median
• non-finite values such as NaN and Inf

Every run writes three local reports:

summary.json   machine-readable results
report.md      portable review document
report.html    visual report for browsers and screenshots

The source CSV is read-only and is never uploaded or changed.

Native interface

On Windows, double-click:

run_gui.bat

The interface lets you select a CSV, choose a report folder, generate the report bundle and open the HTML report. The included synthetic demo is preselected on first launch.

You can also run the installed command:

reportkit-gui

Command line

No third-party runtime dependency is required.

python -m reportkit inspect examples/demo.csv --out demo-report

To print the JSON result as well:

python -m reportkit inspect examples/demo.csv --out demo-report --json

Open demo-report/report.html in a browser after the command completes.

Install for development

python -m venv .venv
.venv\Scripts\python -m pip install -e .
.venv\Scripts\python -m unittest discover -s tests -v

Supported Python versions: 3.10, 3.11 and 3.12.

Safety and privacy

• local processing only
• no network calls
• no source-file writes
• no silent row deletion or missing-value imputation
• no business or scientific conclusions
• HTML content is escaped before rendering
• synthetic example data only in the repository

The report identifies structural signals; a domain owner must decide what they mean and whether any correction is appropriate.

Current scope

Version 0.2.0 supports CSV inspection and report generation. It does not yet support Excel workbooks, automated cleaning, database connections or large-file streaming. These are intentional product boundaries, not implied features.

Tests

python -m unittest discover -s tests -v

GitHub Actions runs the suite on Windows with Python 3.10, 3.11 and 3.12.

Project structure

reportkit/core.py       CSV parsing and quality summary
reportkit/reporting.py  JSON, Markdown and safe HTML reports
reportkit/cli.py        command-line interface
reportkit/gui.py        standalone native interface
tests/                  behavior and regression tests
examples/               synthetic demonstration data

License

MIT © 2026 mmm-05-aaa