@wlearn/types

TypeScript interfaces and constants for wlearn. Zero runtime code. This is the contract that all wlearn packages implement against.

Part of wlearn (GitHub, all packages).

Install

npm install @wlearn/types

Constants

const { BUNDLE_MAGIC, BUNDLE_VERSION, HEADER_SIZE, DTYPE } = require('@wlearn/types')

BUNDLE_MAGIC    // Uint8Array [0x57, 0x4c, 0x52, 0x4e] ('WLRN')
BUNDLE_VERSION  // 1
HEADER_SIZE     // 16 bytes
DTYPE           // { FLOAT32: 'float32', FLOAT64: 'float64', INT32: 'int32' }

Types

Data types:

• DenseMatrix -- { data: Float32Array | Float64Array, rows, cols }
• CSRMatrix -- { data, indices, indptr, rows, cols } (compressed sparse row)
• Matrix -- DenseMatrix | CSRMatrix
• Labels -- Int32Array | Float32Array | Float64Array
• TensorRef -- zero-copy view descriptor for pipeline data routing

Estimator contract:

• Estimator -- fit(), predict(), score(), save(), dispose(), getParams(), setParams()
• Classifier -- extends Estimator with predictProba() and classes
• Transformer -- fit(), transform(), fitTransform(), save(), dispose()
• Capabilities -- runtime feature flags (classifier, predictProba, csr, etc.)

AutoML:

• SearchParam -- hyperparameter distribution (categorical, uniform, log_uniform, int_uniform)
• SearchSpace -- Record<string, SearchParam> for model-provided search spaces
• AutoFitOpts, AutoFitResult -- options and result for autoFit()

Bundle format:

• BundleManifest -- manifest with typeId, params, metadata
• BundleTOCEntry -- { id, offset, length, sha256, mediaType }
• LoaderFn -- (manifest, toc, blobs) => Estimator | Promise<Estimator>

License

Apache-2.0