Skip to content

flurie/sdp

 
 

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Simple Data Processing

It is a library for simple data processing. sdp is inspired by array, vector, bytestring, containers and repa. sdp is focused on efficiency, openness and simplicity.

Motivation

Originally sdp was created as array replacement. Now its main purpose was to provide the most comfortable interface for working with various data structures: eliminate name conflicts with other libraries and between them, simplify their interaction, make code even more versatile and reusable.

Structures

All predefined structures are based on pseudo-primitive types SArray#, SBytes#, STArray#, STBytes#, MIOArray#, MIOBytes# that encapsulate real primitives with correct size and offset.

sdp uses templates to define more complex structures:

  • AnyBorder adds explicit boundaries of arbitrary type
  • AnyChunks defines an unrolled list of given elements

Based on pseudo-primitives and templates, the following are defined:

  • immutable arrays with immutable explicit immutable boundaries: Array and Bytes
  • mutable arrays with immutable explicit immutable boundaries: STArray and STBytes, MIOArray (IOArray) and MIOBytes (IOBytes)
  • immutable unrolled lists with immutable implicit boundaries: Unlist and Ublist
  • mutable unrolled lists with immutable implicit boundaries: STUnlist and STUblist, MIOUnlist (IOUnlist) and MIOUblist (IOUblist)
  • immutable unrolled lists with explicit immutable boundaries: Unrolled and ByteList
  • mutable unrolled lists with explicit immutable boundaries STUnrolled and STByteList, MIOUnrolled (IOUnrolled) and MIOByteList (IOByteList)

Classes

sdp generalize the most popular operations on linear (list-like) and associative data structures including selection, splitting and sorting. With sdp list functions aren't overlap their counterparts for other structures.

  • Nullable is a service class of structures with null values.
  • Bordered is a class of structures with borders and finite number of elements.
  • Estimate is a service class for efficiently (at least, in finite time) estimating the length of a structure or compare pair of structures by length (in finite time if at least one of them is finite). Allows to express such conditions as:
xs .<. ys -- length xs < length ys
es .== 10 -- length es == 10
es .>   5 -- length es > 5
  • Unboxed is a service class that simplifies interacting with data stored in
  • ByteArray# and MutableArray#. Used in containers that based on SBytes#, STBytes# or IOBytes#.
  • Shape is a service class for dimension operations and finite-dimensional index transformations. Index is a service class that generalizes Enum to interval operations, replaces Ix.
  • Linear is a class of linear structures that generalizes the standard list functions. Split is Linear extension, which implements additional list-like operations like split(At) takeWhile, isPrefixOf, etc.
  • Indexed is a class of indexed structures that generalizes read and modify operations immutable structures.
  • Shaped is a class of operations on structures generalized by the type of index. Provides safe change of range and index type, fast extraction of subsequences.
  • Map is a class of operations on associative arrays (dictionaries).
  • Set is a class of operations on sets.
  • Zip is a class that generalizes zipping operations (see zip, zipWith, ZipList, etc.).
  • Scan is a class of convolutions with intermediate values.
  • Sort is a sorting class for immutable structures.
  • BorderedM, LinearM, SplitM, IndexedM, SortM - classes of operations on mutable containers.

Versions

sdp follow Haskell Package Versioning Policy. To simplify the search for extensions, I also recommend the following rules:

  • The MAJOR version of the derivative must match the smallest MAJOR version of sdp with which it's compatible.
  • Extensions should be called sdp-%extensionname% (e.g. sdp-quickcheck).
  • Wrappers should be called sdp4%libraryname% (e.g., sdp4text).

Using the SDP category

The SDP category is intended for sdp classes and primitives, as for structures whose names are already taken in the Data category. It shouldn't be used instead of System, Control, Foreign, etc.

Contributing

For details of the process for submitting pull requests, please read CONTRIBUTING.md.

License

sdp is FOSS (free and open source software), you can redistribute it and/or modify it under the terms of the BSD3 license. sdp is distributed in the hope that it will be useful, but without any warranty, without even the implied warranty of merchantability or fitness for a particular purpose. See the LICENSE file for more details.

About

Simple Data Processing

Resources

License

Stars

Watchers

Forks

Packages

No packages published

Languages

  • Haskell 100.0%