KoreanRomanization is a C# class library for romanizing Korean text.
With this class library, it is possible to take a string of text, which contains some characters from the native Korean script - Hangeul - and convert it into a string of text where the Korean characters have been replaced with a representation of their pronunciation using the English alphabet. This class library can romanize Korean text using any of four systems of romanization (different romanization systems have different advantages and disadvantages in terms of accuracy, readability, writeability, and so on).
The first part of the process is to divide the text that is given into the sections - the Korean text and the non-Korean text. The non-Korean sections are ignored by the romanization (and can either be left in the text as they are, or removed to leave only the romanized Korean). This part of the process is done by the GetTextBlock() function in the base Romanization class. This function takes a string, and returns a TextBlock object, which contains a list of ITextSections, which are either of the type KoreanTextSection for Korean text, or NonKoreanTextSection for everything else.
The RomanizeTextBlock() function within Romanization then takes this TextBlock object, and goes through each section of it. For each KoreanTextSection, it will go through each of the Korean syllables within that section, and romanize each one based on the rules of the romanization system.
There are four derived classes: McCuneReischauerRomanization, RevisedRomanization, YaleRomanization, and SimplifiedRomanization, which each inherit from Romanization, and can romanize text according to the rules of those four systems of romanization. The majority of the rules for each system are defined in a set of five enumerable properties of the class; however, in several of the systems of romanization there are certain rules which apply only under very specific circumstances, and depend on a very specific combination of initials, medials, and finals over several syllables, and so the entirety of the rules cannot be expressed in this way, and finer rules are sometimes just defined as part of the basic romanization functions for each class.
To romanize text according to a specific system, first create an instance of the romanization class for the system you want to use, and then call the RomanizeText() function, which takes a string as its argument, and returns a single string with the romanized text.
The KoreanLetter struct encapsulates a letter in the Korean language - known as a jamo.
It has a number of internal constants that help define which letter a given character is, as well as shortcuts for instantiating particular Korean letters. This is useful for implementing some of the rules of romanization.
The KoreanSyllable struct encapsulates a syllable in the Korean language - which comprises of an initial, a medial, and a final letter, or just an initial and a medial letter. Every possible syllable that exists in Korean is stored as a separate character code in Unicode, but in order to romanize any Korean text, you need to know what the individual letters of the syllable are. The KoreanSyllable struct provides methods for identifying the letters within a syllable.
Represents a generic block of text, which can be described in terms of Korean and non-Korean sections. Any text that is passed through to the romanization converters is first converted into a TextBlock object.
Represents a section of Korean text, which is made up of syllabic blocks.
Represents a section of non-Korean text. This is just a text string, and it is ignored by the romanization converter.
There are two TupleList classes, and they both simply represent a list of tuples.
These simple classes allow for easy hardcoding of romanization data into the class library. The shortest way to express a romanization rule is as a tuple, which requires minimal markup. TupleList allows a large number of tuples to be added to a list with very little markup.
The Romanization class acts as a base class for other specific romanization schemes. It encapsulates the basic properties of a romanization converter, such as whether it will explicitly romanize ㅅ as 'sh' when it is followed by ㅣ. It also contains the functions used to deconstruct a passage of text into its Korean and non-Korean components, as well as core, overridable functions that are used to romanize the text.
A further abstraction of the concept of a romanization converter. There are only three official systems of romanization for Korean, so it is unlikely that anyone would want to write a custom converter - even less so one that didn't implement the Romanization base class. But if they did, the IRomanization describes the basic features of a romanization converter.
The KoreanRomanization library is covered by a large number of unit tests - 1239 of them.