
Generate 'credits' for included code: parses SPDX licenses, groups dependencies
by license, groups licensed dependencies by author. In the interest of creating
concise credits pages.
- Detects license type from the
license specification in package.json, which
is parsed as SPDX
- Detects authorship from, in order of preference:
author field in package.json
authors field in package.json
maintainers field in package.json
- Parsing a license file, one of
- LICENSE
- LICENSE.md
- LICENSE.txt
- license
- license.md
- license.txt
Instead of a code bundle, this produces a bundle of credits output. Typically
you'll run this to create a source file that powers your 'credits' page. The
format of this output is as an ES module with a default export. The export is
a list of licenses with { license, modules } keys, which contain a list
of authors with { author, modules } keys.
export default [
{
license: { license: "MIT" },
modules: [
{
author: "James Halliday",
modules: [
"resumer",
"object-inspect"
]
},
{ author: "Dominic Tarr", modules: ["through"] },
{
author: "Marijn Haverbeke and Ingvar Stepanyan",
modules: ["acorn"]
}
]
},
{
license: { license: "ISC" },
modules: [{ author: "Isaac Z. Schlueter", modules: ["inherits"] }]
}
];
- whitelist: an array or Set of SPDX license identifiers that are allowed in the
list. If provided and a non-whitelisted dependency is encountered, the transform
will reject.