A molecule parser written in JavaScript that helps developers to parse molecule into a codec map.

const { createParser } = require("@ckb-lumos/molecule");

const parser = createParser();
const codecMap = parser.parse(`
  array Uint8 [byte; 1];
`);

codecMap.Uint8.pack(1);

A CLI to generate a set of TypeScript-friendly codec from a Molecule schema file. To use it, you need to install @ckb-lumos/molecule and @ckb-lumos/codec first. The @ckb-lumos/molecule could be devDependencies if you only use it for codegen.

npm install -D @ckb-lumos/molecule
npm install @ckb-lumos/codec

Then you can create a lumos-molecule-codegen.json file to configure the codegen.

Note: if you find the npx lumos-molecule-codegen command is not found, please try to replace npx with node_modules/.bin/lumos-molecule-codegen.

// lumos-molecule-codegen.json
{
  // keep | camelcase
  objectKeyFormat: "camelcase",
  // prepend the import statement to custom and override the generated codec
  prepend: "import { Uint32, Uint64, Uint128 } from './customized'",
  // the input schema file
  schemaFile: "blockchain.mol",
}

Finally, run the following command to generate code to write to generated.ts.

npx lumos-molecule-codegen > generated.ts

To generate all the molecule files in a directory that match the pattern of **/*.mol, you can use the following configuration.

{
  // keep | camelcase
  objectKeyFormat: "camelcase",
  // prepend the import statement to custom and override the generated codec
  // to make the relative import work, you need to run the command in the same directory as the `customized` directory
  prepend: "import { Uint32, Uint64, Uint128 } from './customized'",
  // the input schema directory, all **/*.mol in the directory will be processed
  schemaDir: "schemas",
  // the output directory
  outDir: "generated",
}

Then run the following command to generate code to the generated directory.

npx lumos-molecule-codegen

The parser inside @ckb-lumos/molecule is based on the EBNF, but the Rust implementation is based on the Pest, there are some differences between them, such as

• comment starts with #
• comment defined in the struct, table, or union is not supported well

Therefore, to use lumos-molecule-codegen, you need to make sure the comments should not be placed