tokenizer-properties.js

define([
	"./get-time-zone-name",
	"./pattern-re",
	"./timezone-hour-format/h",
	"./timezone-hour-format/hm",
	"../common/create-error/unsupported-feature",
	"../common/format-message",
	"../number/numbering-system-digits-map",
	"../number/symbol",
	"../util/is-plain-object",
	"../util/loose-matching",
	"../util/object/filter",
	"../util/regexp/escape"
], function( dateGetTimeZoneName, datePatternRe, dateTimezoneHourFormatH, dateTimezoneHourFormatHm,
	createErrorUnsupportedFeature, formatMessage, numberNumberingSystemDigitsMap, numberSymbol,
	isPlainObject, looseMatching, objectFilter, regexpEscape ) {

/**
 * tokenizerProperties( pattern, cldr )
 *
 * @pattern [String] raw pattern.
 *
 * @cldr [Cldr instance].
 *
 * Return Object with data that will be used by tokenizer.
 */
return function( pattern, cldr, timeZone ) {
	var digitsReSource,
		properties = {
			pattern: looseMatching( pattern )
		},
		timeSeparator = numberSymbol( "timeSeparator", cldr ),
		widths = [ "abbreviated", "wide", "narrow" ];

	digitsReSource = numberNumberingSystemDigitsMap( cldr );
	digitsReSource = digitsReSource ? "[" + digitsReSource + "]" : "\\d";
	properties.digitsRe = new RegExp( digitsReSource );

	// Transform:
	// - "+H;-H" -> /\+(\d\d?)|-(\d\d?)/
	// - "+HH;-HH" -> /\+(\d\d)|-(\d\d)/
	// - "+HHmm;-HHmm" -> /\+(\d\d)(\d\d)|-(\d\d)(\d\d)/
	// - "+HH:mm;-HH:mm" -> /\+(\d\d):(\d\d)|-(\d\d):(\d\d)/
	//
	// If gmtFormat is GMT{0}, the regexp must fill {0} in each side, e.g.:
	// - "+H;-H" -> /GMT\+(\d\d?)|GMT-(\d\d?)/
	function hourFormatRe( hourFormat, gmtFormat, digitsReSource, timeSeparator ) {
		var re;

		if ( !digitsReSource ) {
			digitsReSource = "\\d";
		}
		if ( !gmtFormat ) {
			gmtFormat = "{0}";
		}

		re = hourFormat
			.replace( "+", "\\+" )

			// Unicode equivalent to (\\d\\d)
			.replace( /HH|mm|ss/g, "((" + digitsReSource + "){2})" )

			// Unicode equivalent to (\\d\\d?)
			.replace( /H|m/g, "((" + digitsReSource + "){1,2})" );

		if ( timeSeparator ) {
			re = re.replace( /:/g, timeSeparator );
		}

		re = re.split( ";" ).map(function( part ) {
			return gmtFormat.replace( "{0}", part );
		}).join( "|" );

		return new RegExp( "^" + re );
	}

	function populateProperties( path, value ) {

		// Skip
		var skipRe = /(timeZoneNames\/zone|supplemental\/metaZones|timeZoneNames\/metazone|timeZoneNames\/regionFormat|timeZoneNames\/gmtFormat)/;
		if ( skipRe.test( path ) ) {
			return;
		}

		if ( !value ) {
			return;
		}

		// The `dates` and `calendars` trim's purpose is to reduce properties' key size only.
		path = path.replace( /^.*\/dates\//, "" ).replace( /calendars\//, "" );

		// Specific filter for "gregorian/dayPeriods/format/wide".
		if ( path === "gregorian/dayPeriods/format/wide" ) {
			value = objectFilter( value, /^am|^pm/ );
		}

		// Transform object into array of pairs [key, /value/], sort by desc value length.
		if ( isPlainObject( value ) ) {
			value = Object.keys( value ).map(function( key ) {
				return [ key, new RegExp( "^" + regexpEscape( looseMatching( value[ key ] ) ) ) ];
			}).sort(function( a, b ) {
				return b[ 1 ].source.length - a[ 1 ].source.length;
			});

		// If typeof value === "string".
		} else {
			value = looseMatching( value );
		}
		properties[ path ] = value;
	}

	function regexpSourceSomeTerm( terms ) {
		return "(" + terms.filter(function( item ) {
			return item;
		}).reduce(function( memo, item ) {
			return memo + "|" + item;
		}) + ")";
	}

	cldr.on( "get", populateProperties );

	pattern.match( datePatternRe ).forEach(function( current ) {
		var aux, chr, daylightTzName, gmtFormat, length, standardTzName;

		chr = current.charAt( 0 );
		length = current.length;

		if ( chr === "Z" ) {
			if ( length < 5 ) {
				chr = "O";
				length = 4;
			} else {
				chr = "X";
				length = 5;
			}
		}

		// z...zzz: "{shortRegion}", eg. "PST" or "PDT".
		// zzzz: "{regionName} {Standard Time}" or "{regionName} {Daylight Time}",
		//       e.g., "Pacific Standard Time" or "Pacific Daylight Time".
		// http://unicode.org/reports/tr35/tr35-dates.html#Date_Format_Patterns
		if ( chr === "z" ) {
			standardTzName = dateGetTimeZoneName( length, "standard", timeZone, cldr );
			daylightTzName = dateGetTimeZoneName( length, "daylight", timeZone, cldr );
			if ( standardTzName ) {
				standardTzName = regexpEscape( looseMatching( standardTzName ) );
			}
			if ( daylightTzName ) {
				daylightTzName = regexpEscape( looseMatching( daylightTzName ) );
			}
			if ( standardTzName || daylightTzName ) {
				properties.standardOrDaylightTzName = new RegExp(
					"^" + regexpSourceSomeTerm([ standardTzName, daylightTzName ])
				);
			}

			// Fall through the "O" format in case one name is missing.
			if ( !standardTzName || !daylightTzName ) {
				chr = "O";
				if ( length < 4 ) {
					length = 1;
				}
			}
		}

		// v...vvv: "{shortRegion}", eg. "PT".
		// vvvv: "{regionName} {Time}" or "{regionName} {Time}",
		// e.g., "Pacific Time"
		// http://unicode.org/reports/tr35/tr35-dates.html#Date_Format_Patterns
		if ( chr === "v" ) {
			if ( length !== 1 && length !== 4 ) {
				throw createErrorUnsupportedFeature({
					feature: "timezone pattern `" + pattern + "`"
				});
			}
			var genericTzName = dateGetTimeZoneName( length, "generic", timeZone, cldr );
			if ( genericTzName ) {
				properties.genericTzName = new RegExp(
					"^" + regexpEscape( looseMatching( genericTzName ) )
				);
				chr = "O";

			// Fall back to "V" format.
			} else {
				chr = "V";
				length = 4;
			}
		}

		switch ( chr ) {

			// Era
			case "G":
				cldr.main([
					"dates/calendars/gregorian/eras",
					length <= 3 ? "eraAbbr" : ( length === 4 ? "eraNames" : "eraNarrow" )
				]);
				break;

			// Year
			case "u": // Extended year. Need to be implemented.
			case "U": // Cyclic year name. Need to be implemented.
				throw createErrorUnsupportedFeature({
					feature: "year pattern `" + chr + "`"
				});

			// Quarter
			case "Q":
			case "q":
				if ( length > 2 ) {
					cldr.main([
						"dates/calendars/gregorian/quarters",
						chr === "Q" ? "format" : "stand-alone",
						widths[ length - 3 ]
					]);
				}
				break;

			// Month
			case "M":
			case "L":

				// number l=1:{1,2}, l=2:{2}.
				// lookup l=3...
				if ( length > 2 ) {
					cldr.main([
						"dates/calendars/gregorian/months",
						chr === "M" ? "format" : "stand-alone",
						widths[ length - 3 ]
					]);
				}
				break;

			// Day
			case "g":

				// Modified Julian day. Need to be implemented.
				throw createErrorUnsupportedFeature({
					feature: "Julian day pattern `g`"
				});

			// Week day
			case "e":
			case "c":

				// lookup for length >=3.
				if ( length <= 2 ) {
					break;
				}

			/* falls through */
			case "E":
				if ( length === 6 ) {

					// Note: if short day names are not explicitly specified, abbreviated day
					// names are used instead http://www.unicode.org/reports/tr35/tr35-dates.html#months_days_quarters_eras
					// eslint-disable-next-line no-unused-expressions
					cldr.main([
						"dates/calendars/gregorian/days",
						[ chr === "c" ? "stand-alone" : "format" ],
						"short"
					]) || cldr.main([
						"dates/calendars/gregorian/days",
						[ chr === "c" ? "stand-alone" : "format" ],
						"abbreviated"
					]);
				} else {
					cldr.main([
						"dates/calendars/gregorian/days",
						[ chr === "c" ? "stand-alone" : "format" ],
						widths[ length < 3 ? 0 : length - 3 ]
					]);
				}
				break;

			// Period (AM or PM)
			case "a":
				cldr.main(
					"dates/calendars/gregorian/dayPeriods/format/wide"
				);
				break;

			// Zone
			case "V":

				if ( length === 1 ) {
					throw createErrorUnsupportedFeature({
						feature: "timezone pattern `" + pattern + "`"
					});
				}

				if ( timeZone ) {
					if ( length === 2 ) {

						// Skip looseMatching processing since timeZone is a canonical posix value.
						properties.timeZoneName = timeZone;
						properties.timeZoneNameRe = new RegExp( "^" + regexpEscape( timeZone ) );
						break;
					}

					var timeZoneName,
						exemplarCity = cldr.main([
							"dates/timeZoneNames/zone", timeZone, "exemplarCity"
						]);

					if ( length === 3 ) {
						if ( !exemplarCity ) {
							exemplarCity = cldr.main([
								"dates/timeZoneNames/zone/Etc/Unknown/exemplarCity"
							]);
						}
						timeZoneName = exemplarCity;
					}

					if ( exemplarCity && length === 4 ) {
						timeZoneName = formatMessage(
							cldr.main(
								"dates/timeZoneNames/regionFormat"
							),
							[ exemplarCity ]
						);
					}

					if ( timeZoneName ) {
						timeZoneName = looseMatching( timeZoneName );
						properties.timeZoneName = timeZoneName;
						properties.timeZoneNameRe = new RegExp(
							"^" + regexpEscape( timeZoneName )
						);
					}
				}

				if ( current === "v" ) {
					length = 1;
				}

			/* falls through */
			case "z":
			case "O":
				gmtFormat = cldr.main( "dates/timeZoneNames/gmtFormat" );
				cldr.main( "dates/timeZoneNames/gmtZeroFormat" );
				cldr.main( "dates/timeZoneNames/hourFormat" );
				properties[ "timeZoneNames/gmtZeroFormatRe" ] =
					new RegExp( "^" + regexpEscape( properties[ "timeZoneNames/gmtZeroFormat" ] ) );
				aux = properties[ "timeZoneNames/hourFormat" ];
				properties[ "timeZoneNames/hourFormat" ] = (
					length < 4 ?
						[ dateTimezoneHourFormatHm( aux, "H" ), dateTimezoneHourFormatH( aux ) ] :
						[ dateTimezoneHourFormatHm( aux, "HH" ) ]
				).map(function( hourFormat ) {
					return hourFormatRe(
						hourFormat,
						gmtFormat,
						digitsReSource,
						timeSeparator
					);
				});

			/* falls through */
			case "X":
			case "x":

				// x: hourFormat("+HH[mm];-HH[mm]")
				// xx: hourFormat("+HHmm;-HHmm")
				// xxx: hourFormat("+HH:mm;-HH:mm")
				// xxxx: hourFormat("+HHmm[ss];-HHmm[ss]")
				// xxxxx: hourFormat("+HH:mm[:ss];-HH:mm[:ss]")
				properties.x = [
					[ "+HHmm;-HHmm", "+HH;-HH" ],
					[ "+HHmm;-HHmm" ],
					[ "+HH:mm;-HH:mm" ],
					[ "+HHmmss;-HHmmss", "+HHmm;-HHmm" ],
					[ "+HH:mm:ss;-HH:mm:ss", "+HH:mm;-HH:mm" ]
				][ length - 1 ].map(function( hourFormat ) {
					return hourFormatRe( hourFormat );
				});
		}
	});

	cldr.off( "get", populateProperties );

	return properties;
};

});