/// <summary>
/// Removes the specified vishraams from a string.
/// </summary>
/// <param name="text">The text to remove vishraams from.</param>
/// <param name="vishraams">The vishraams to remove. Defaults to all.</param>
/// <returns>A vishraam-less Gurmukhi string.</returns>
public static string StripVishraams(string text,
Vishraam vishraams = Vishraam.Heavy | Vishraam.Medium | Vishraam.Light)
{
var vishraamMapping = Mapping.VishraamMapping();
var vishraamOptions = Enum.GetValues(vishraams.GetType())
.Cast <Enum>()
.Where(vishraams.HasFlag)
.Select(value => vishraamMapping[value.ToString()].ToString());
return(Regex.Replace(text, StringRegexHelper.GetRegexClass(vishraamOptions), ""));
}
/// <summary>
/// Strips line endings from any Gurmukhi or translation string.
/// Accepts both Unicode and ASCII input.
/// Useful for generating accurate first letters or modifying non-Gurbani for better display.
/// *Not* designed for headings or Sirlekhs.
/// </summary>
/// <param name="text">The text to stip endings from.</param>
/// <returns>A ending-less version of the text.</returns>
public static string StripEndings(string text)
{
// Line endings in both ASCII, Unicode, and English
var endingClass = StringRegexHelper.GetRegexClass(new[] { "।", "॥", "]", "[", "|" });
// Sometimes translation line endings begin with these characters, before numbers
var optionalEndingClass = StringRegexHelper.GetRegexClass(new[] { "(" });
// Remove any broken endings
var brokenEndingClass = StringRegexHelper.GetRegexGroup(new[] { "()" });
// All numbers in ASCII, Unicode
var numbers = Enumerable.Range(0, 10).ToArray()
.Select(i => i.ToString())
.ToArray();
var numberClass = StringRegexHelper.GetRegexClass(
numbers
.Concat(numbers.Select(ToUnicodeGurmukhi))
.Concat(numbers.Select(ToUnicodeGurmukhi).Select(ToHindi))
);
// Rahao in English, ASCII, Unicode
var pauseGroup = StringRegexHelper
.GetRegexGroup(new[] { "ਰਹਾਉ", ToAsciiGurmukhi("ਰਹਾਉ"), "Pause" });
// Matchers to strip out of input string
var matchers = new[]
{
// Any pause (ending + pause word) => match the rest of the line
$" ?{endingClass} ?{pauseGroup}.*",
// Any ending followed by any number => match the rest of the line
$" ?({endingClass}|{optionalEndingClass}){numberClass}.*",
// Any sequence at the end of a line with numbers, periods, and spaces beginning with a number
$" ?{numberClass}({numberClass}|[. ])*$",
// Clean up any lingering ending characters
$" ?{brokenEndingClass}",
$" ?{endingClass}",
};
return(matchers.Aggregate(text, (a, b) =>
Regex.Replace(a, b, "").Trim()));
}
/// <summary>
/// Represents text in syllables according to Sanskrit prosody, Pingala, Matra/Meter/Morae
/// </summary>
/// <param name="text">The string to convert</param>
/// <returns>A syllabic representation of 1"s (laghu/light/short) and 2"s (guru/heavy/long).</returns>
public static string ToSyllabicSymbols(string text)
{
// These have no impact on weight, therfore remove before processing
var zeroWeightSigns = new[]
{
"੍ਰ",
"ੵ",
"ੑ",
"੍ਵ",
"੍ਹ",
"੍ਟ",
"੍ਨ",
"੍ਯ",
"੍ਚ",
"੍ਤ",
"੍",
"ਿ",
"ੁ",
"ਂ",
"ਃ",
"।",
"॥",
"☬",
"਼",
"❁",
};
// For the rest, need to analyze the string
// Base characters counts as one syllable (light)
// Base characters with any number of long (deeragh) sounds counts as two syllables (heavy)
var syllableSymbols = new
{
Light = "1",
Deeragh = "s",
HeavySequence = "1s",
Heavy = "2",
};
// The list of base characters for analysis
var baseCharacters = new[]
{
"ਇ",
"ਉ",
"ਙ",
"ੳ",
"ਅ",
"ਬ",
"ਭ",
"ਚ",
"ਛ",
"ਦ",
"ਧ",
"ੲ",
"ਡ",
"ਢ",
"ਗ",
"ਘ",
"ਹ",
"ਜ",
"ਝ",
"ਕ",
"ਖ",
"ਲ",
"ਲ਼",
"ਮ",
"ਨ",
"ਪ",
"ਫ",
"ਤ",
"ਥ",
"ਰ",
"ਸ",
"ਸ਼",
"ਟ",
"ਠ",
"ਵ",
"ੜ",
"ਣ",
"ਯ",
"ਜ਼",
"ਗ਼",
"ਖ਼",
"ਫ਼",
"ਞ",
};
// The list of long sound characters for analysis
var deeraghModifiers = new[]
{
"ੰ",
"੍ਹੂ",
"ੀ",
"ੂ",
"ੇ",
"ੈ",
"ੋ",
"ੌ",
"ਾਂ",
"ਾ",
"ੱ",
};
// Some symbols represent multiple characters in a single unicode entity.
// These should be considered unsplittable and unmodifiable by other vowels/signs.
// The following represent a base character and deeragh in one unicode entity point.
// Since author is unaware of whether these can be further modified with deeragh
// they"ll be processed as a heavy sequence (base char + deeragh) in the mapping.
// If that is not the case and these characters cannot have further deeragh modifiers,
// then they can be safely mapped directly to syllableSymbol.heavy
var twoUnitSyllables = new[] { "ਊ", "ਓ", "ਈ", "ਏ", "ਐ", "ਆ", "ਔ" };
// Create a map for each character to a syllableSymbol, for further processing/analysis
var syllabicMapping = new[]
{
new { Symbol = syllableSymbols.Light, GroupedCharacters = baseCharacters },
new { Symbol = syllableSymbols.Deeragh, GroupedCharacters = deeraghModifiers },
new { Symbol = syllableSymbols.HeavySequence, GroupedCharacters = twoUnitSyllables }
}
.SelectMany(sm =>
sm.GroupedCharacters.Select(c => new { Character = c, sm.Symbol }))
// Add any missing mappings which do not fit in the above patterns
// These rules are "as is", and are not modified in the reducer above
.Append(new { Character = "ੴ", Symbol = "21 2221" }) // Ik Oankaar / ਇੱਕ ਓਅੰਕਾਰ
.Append(new { Character = " ", Symbol = " " }) // Preserve spacing between words
.ToDictionary(k => k.Character, v => v.Symbol);
// Create Regex rules for replacements
var zeroWeightSignsRegex = new Regex(StringRegexHelper.GetRegexGroup(zeroWeightSigns));
var multipleSpaceCharsRegex = new Regex(" +");
var multipleDeeraghSymbolsRegex = new Regex("s+");
var heavySequenceRegex = new Regex("1s");
text = ToUnicodeGurmukhi(text);
var result = string
.Join("", zeroWeightSignsRegex
.Replace(text, "")
.ToCharArray()
.Select(value =>
syllabicMapping.TryGetValue(value.ToString(), out var val) ? val : " "));
result = multipleSpaceCharsRegex.Replace(result, " ");
result = multipleDeeraghSymbolsRegex.Replace(result, syllableSymbols.Deeragh);
result = heavySequenceRegex.Replace(result, syllableSymbols.Heavy);
return(result);
}
