Skip to main content

A multi-dimensional grid used for state space searching

Project description

An iterable multi-dimensional grid used for exhaustive search and dimensional problem solving.

Description

Defines a Grid object which allows for efficient iteration across an arbitrary number of dimensions. The grid objects allow for iteration, multi-dimensional select, and multi-dimensional slicing.

Accessing a grid follows the __getitem__ convention of [index]. To retrieve a multi-dimensional selection either [index1, index2] or [(index1, index2)] may be used to specify the ordered dimensions of the grid to subselect. Any of the index requests can be replaced by a slice object such that [index1, start2:end2:step2] is a legal request.

Slices provide SubGrid objects which act as grids, but map their referenced data back to the original grid object. This allows for repeated slicing of a grid with near constant memory overhead at the cost of layered slice requests for each change on the original data.

There are several provided Grids which are setup for efficiency for the given data type. So far those include IntGrid, FloatGrid, and ObjectGrid – the latter of which is a general solution without efficient storage. These grid types define the data being stored, rather than the indexing scheme. All grids use an integer based indexing, though there are plans to create a float range grid which does the float to index mapping behind the interface.

Note that creating grids with many dimensions can take up an extremely large amount of memory, even when using an efficient scheme. This extends to a very long iteration times as the number of elements to visit grows exponentially. Take a 5 dimensional grid with 10 values for each dimension. this makes a 10^5 element grid – which is 100k iterables – and would take ~400kb of storage space. The same grid with 100 values for each dimension would have 40 billion elements and take more than 37GB of memory to store.

Dependencies

  • pydatawrap

  • numpy

Setup

Installation

From source:

python settup.py install

From pip:

pip install gridwalker

Features

  • Multi-Dimensional grid definitions with arbitrary number of dimensions

  • Iteration and assignment through efficient means across any dimensional assignment

Language Preferences

  • Google Style Guide

  • Object Oriented (with a few exceptions)

TODO

  • Create float index grids for floating precision grid iteration

Author

Author(s): Matthew Seal

© Copyright 2013, OpenGov

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

GridWalker-1.0.1.zip (13.7 kB view hashes)

Uploaded Source

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page