npy library

Classes

ByteTransformer: Transforms a stream to emit chunks of the specified bufferSize. If bufferSize is not provided, the stream will be emitted as chunks of default size.
NdArray<T>
NpyDType
NpyHeader<T>
NpyHeaderSection
NpyParser<T>
NpyVersion: The version of the numpy file. It is composed of a major and minor version. Supported major version are currently 1, 2 and 3. Supported minor version are currently 0.
NpzFile: The NpzFile class represents a zip file containing one or more .NPY files. It provides methods for loading, saving, and manipulating the contents of the .NPZ file.

NpyEndian: The endianness of the NPY file. It is represented by a single character. Single-byte data types are always NpyEndian.none.
NpyType: The supported data types of the NPY file. A data type is represented by one or multiple single-character representations. If more than one representation exists, the first one is used for saving. This package aims to gradually increase support for more data types.

magicString → const String: Marks the beginning of an NPY file.
newLineOffset → const int: The header section is terminated with a single byte that is a newline character.
supportedInputTypes → const Set<Type>: The supported input types for the NdArray class.

flattenCOrder<T>(List list) → List<T>
flattenFortranOrder<T>(List list, {required List<int> shape}) → List<T>
getPaddingSize(int size) → int: Returns the number of padding bytes needed to make the given size a multiple of 64.
littleEndian16ToInt(List<int> bytes) → int: Converts the given bytes to a 16-bit unsigned integer in little-endian byte order.
littleEndian32ToInt(List<int> bytes) → int: Converts the given bytes to a 32-bit unsigned integer in little-endian byte order.
parseByteData<T>(List<int> bytes, NpyDType dtype) → List<T>: Parse byte data according to the dtype and return a one-dimensional, flattened List of values.
reshape<T>(List<T> oneDimensionalList, List<int> shape, {bool fortranOrder = false}) → List: Reshape a one-dimensional List according to the given shape and order (C or Fortran).
save(String path, List list, {NpyDType? dtype, NpyEndian? endian, bool? fortranOrder}) → Future<void>: Convenience function to save a List to the given path in NPY format. Alternatively use NdArray.save.