Skip to content

Add MIT licence and a notice separating code from data - #1

Open
DataScienceVishal wants to merge 1 commit into
mainfrom
add-licence
Open

Add MIT licence and a notice separating code from data#1
DataScienceVishal wants to merge 1 commit into
mainfrom
add-licence

Conversation

@DataScienceVishal

Copy link
Copy Markdown
Owner

Why

This repository had no LICENSE file. Public on GitHub is not a licence, so
default copyright applied: readers could look at the code and legally could not
reuse, fork or adapt any of it.

What changed

Adds the standard MIT licence, Copyright (c) 2026 Vishal Khan.

It also adds NOTICE.md, because on this repository an MIT file on its own is
actively misleading. This project deliberately commits no HAM10000 images and
fetches them at runtime instead, so a reader who sees only LICENSE could
reasonably conclude that MIT covers everything the project touches, including
the data. It does not, and the data is non-commercial only.

NOTICE.md states the split directly:

  • MIT covers the code in this repository
  • the HAM10000 corpus that uv run lesion-split fetch downloads is
    CC BY-NC 4.0, non-commercial use only, Harvard Dataverse version 4.0,
    doi:10.7910/DVN/DBW86T, credited to Tschandl, Rosendahl and Kittler,
    Scientific Data 5, 180161 (2018)
  • the MIT grant does not extend to the data, and the non-commercial restriction
    on the data does not restrict the code

The README change is two lines

The README's data section already cited the dataset licence, the DOI and the
paper, and it already explained why nothing is committed. What it never said was
what licence the code itself carries. It gains two lines pointing at
NOTICE.md, inserted after the existing paragraph on non-commercial terms.

No figure, number, table, marker or existing sentence in the README was
modified. The diff there is three added lines and zero removed, which the
generated-table checks confirm.

Tests

uv sync --all-extras && uv run pytest on the branch:

70 passed in 3.81s

Zero failures, zero errors, zero skips. This matters slightly more here than for
a normal licence commit, because tests/test_tables.py fails when the README
stops matching the committed artifacts, so a green suite is evidence the README
addition did not disturb any generated figure.

🤖 Generated with Claude Code

This repository had no LICENSE file, so default copyright applied and
nobody could legally reuse the code regardless of it being public.

Adds the standard MIT licence, copyright 2026 Vishal Khan.

Because this project deliberately commits no HAM10000 images and fetches
them at runtime instead, an MIT file on its own could be read as
covering the dataset too. NOTICE.md states the split directly: MIT
covers the code in this repository, and the HAM10000 corpus the fetch
step downloads is CC BY-NC 4.0, non-commercial only, credited to
Tschandl, Rosendahl and Kittler, Harvard Dataverse
doi:10.7910/DVN/DBW86T.

The README already cited the dataset licence and the paper in its data
section but said nothing about the code's own terms. It gains two lines
pointing at NOTICE.md. No existing figure or sentence there was changed.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant