Skip to content

Latest commit

 

History

History
executable file
·
95 lines (76 loc) · 4.08 KB

README.md

File metadata and controls

executable file
·
95 lines (76 loc) · 4.08 KB
Raw

BLDSuite

A collection of useful Java code related to BLD solving

Setup

There is no actual setup involved because this is just some code. You will however need to download several libraries to get this project running

Dependencies

A local MySQL server setup for database testing is recommended (although not necessary)

Modules

The entire suite consists of smaller sub-units, so called "modules". Every module has its own functionality, for which the base workings will be explained below.

The plan is to have all small sub-units integrate as gradle or maven module at some point in the future.

GUIs will also be written and published here once available!

model

Contains mostly enums to facilitate work with cube-related properties. Most importantly, this contains the piece package which is heavily used throughout the entire suite.

Algorithms for the WCA-official BLD events can all be accessed via the CubicPieceType.* enum constants

database

A simple backend for managing persistent algorithm storage. Currently needs a custom DDL to be set up:

CREATE TABLE Algorithms(
	type VARCHAR(255),
	`case` VARCHAR(2),
	alg VARCHAR(255),
	score FLOAT DEFAULT 0,
	review BOOLEAN DEFAULT FALSE,
	buffer VARCHAR(255),
	UNIQUE KEY (type, `case`, alg, buffer)
) ENGINE=InnoDB DEFAULT CHARSET=utf8;
CREATE TABLE Images(
	`case` VARCHAR(2),
	image VARCHAR(255),
	score INT DEFAULT 0,
	token VARCHAR(255),
	`language` VARCHAR(255),
	UNIQUE KEY (`case`, image, token, `language`)
) ENGINE=InnoDB DEFAULT CHARSET=utf8;

analyze

A package for, well, analyzing a given scramble.

Useful for generating memos, evaluating scrambles and reconstructing solves. Can even be provided with an AlgSource (currently only implemented in the database module) to print full algorithm solutions.

Abstract base implementation is defined in GPuzzle.kt, which should be used as static variable type. Puzzles are loaded with file specifications similar to KSolve.

filter

A package for conditional BLD scrambling.

Uses an underlying TNoodle implementation to "brute force" a desired scramble. Please note that depending on your wishes, this can take a very (very, very) long time to yield results.

The abstract base implementation is given in BldScramble.java. Concrete implementations can be found in ThreeBldScramble.java, FourBldScramble.java, FiveBldScramble.java and so on.

The condition sub-package contains the required filtering mechanisms. The following static constructors are available:

IntCondition.java:

  • IntCondition EXACT(int n) specifies exactly n
  • IntCondition INTERVAL(int min, int max) specifies a range between min and max, inclusive
  • IntCondition MIN(int min) specifies at least min
  • IntCondition MAX(int max) specifies at most max
  • IntCondition ANY() specifies any amount (inside the Java 32bit Integer domain)
  • IntCondition NONE() specifies no match

BooleanCondition.java:

  • BooleanCondition YES() specifies a condition must apply
  • BooleanCondition NO() specifies a condition must not apply
  • BooleanCondition MAYBE() specifies a condition may apply but it essentially doesn't matter

Additional filters can be specified via regular expressions, namely:

  • setMemoRegex(PieceType type, String regExp) requires that the given piece type memo matches the given RegExp. No boundary conditions apply, so if you enter silly stuff it might not work
  • setLetterPairRegex(PieceType type, List<String> pairs) requires that the given piece type memo contains all of the letter pairs specified in the list passed as second argument
  • setPredicateRegex(PieceType type, AlgSource algSource, Predicate<Algorithm> filter) requires that the given piece type execution matches the given predicate from start to end, using the algorithms present in the given source