import { constructFrom } from "./constructFrom.mjs";
import { getDefaultOptions } from "./getDefaultOptions.mjs";
import { defaultLocale } from "./_lib/defaultLocale.mjs";
import { toDate } from "./toDate.mjs";
import { longFormatters } from "./_lib/format/longFormatters.mjs";
import {
isProtectedDayOfYearToken,
isProtectedWeekYearToken,
warnOrThrowProtectedError,
} from "./_lib/protectedTokens.mjs";
import { parsers } from "./parse/_lib/parsers.mjs";
import { DateToSystemTimezoneSetter } from "./parse/_lib/Setter.mjs";
// Rexports of internal for libraries to use.
// See: https://github.com/date-fns/date-fns/issues/3638#issuecomment-1877082874
export { longFormatters, parsers };
/**
* The {@link parse} function options.
*/
// This RegExp consists of three parts separated by `|`:
// - [yYQqMLwIdDecihHKkms]o matches any available ordinal number token
// (one of the certain letters followed by `o`)
// - (\w)\1* matches any sequences of the same letter
// - '' matches two quote characters in a row
// - '(''|[^'])+('|$) matches anything surrounded by two quote characters ('),
// except a single quote symbol, which ends the sequence.
// Two quote characters do not end the sequence.
// If there is no matching single quote
// then the sequence will continue until the end of the string.
// - . matches any single character unmatched by previous parts of the RegExps
const formattingTokensRegExp =
/[yYQqMLwIdDecihHKkms]o|(\w)\1*|''|'(''|[^'])+('|$)|./g;
// This RegExp catches symbols escaped by quotes, and also
// sequences of symbols P, p, and the combinations like `PPPPPPPppppp`
const longFormattingTokensRegExp = /P+p+|P+|p+|''|'(''|[^'])+('|$)|./g;
const escapedStringRegExp = /^'([^]*?)'?$/;
const doubleQuoteRegExp = /''/g;
const notWhitespaceRegExp = /\S/;
const unescapedLatinCharacterRegExp = /[a-zA-Z]/;
/**
* @name parse
* @category Common Helpers
* @summary Parse the date.
*
* @description
* Return the date parsed from string using the given format string.
*
* > ⚠️ Please note that the `format` tokens differ from Moment.js and other libraries.
* > See: https://github.com/date-fns/date-fns/blob/master/docs/unicodeTokens.md
*
* The characters in the format string wrapped between two single quotes characters (') are escaped.
* Two single quotes in a row, whether inside or outside a quoted sequence, represent a 'real' single quote.
*
* Format of the format string is based on Unicode Technical Standard #35:
* https://www.unicode.org/reports/tr35/tr35-dates.html#Date_Field_Symbol_Table
* with a few additions (see note 5 below the table).
*
* Not all tokens are compatible. Combinations that don't make sense or could lead to bugs are prohibited
* and will throw `RangeError`. For example usage of 24-hour format token with AM/PM token will throw an exception:
*
* ```javascript
* parse('23 AM', 'HH a', new Date())
* //=> RangeError: The format string mustn't contain `HH` and `a` at the same time
* ```
*
* See the compatibility table: https://docs.google.com/spreadsheets/d/e/2PACX-1vQOPU3xUhplll6dyoMmVUXHKl_8CRDs6_ueLmex3SoqwhuolkuN3O05l4rqx5h1dKX8eb46Ul-CCSrq/pubhtml?gid=0&single=true
*
* Accepted format string patterns:
* | Unit |Prior| Pattern | Result examples | Notes |
* |---------------------------------|-----|---------|-----------------------------------|-------|
* | Era | 140 | G..GGG | AD, BC | |
* | | | GGGG | Anno Domini, Before Christ | 2 |
* | | | GGGGG | A, B | |
* | Calendar year | 130 | y | 44, 1, 1900, 2017, 9999 | 4 |
* | | | yo | 44th, 1st, 1900th, 9999999th | 4,5 |
* | | | yy | 44, 01, 00, 17 | 4 |
* | | | yyy | 044, 001, 123, 999 | 4 |
* | | | yyyy | 0044, 0001, 1900, 2017 | 4 |
* | | | yyyyy | ... | 2,4 |
* | Local week-numbering year | 130 | Y | 44, 1, 1900, 2017, 9000 | 4 |
* | | | Yo | 44th, 1st, 1900th, 9999999th | 4,5 |
* | | | YY | 44, 01, 00, 17 | 4,6 |
* | | | YYY | 044, 001, 123, 999 | 4 |
* | | | YYYY | 0044, 0001, 1900, 2017 | 4,6 |
* | | | YYYYY | ... | 2,4 |
* | ISO week-numbering year | 130 | R | -43, 1, 1900, 2017, 9999, -9999 | 4,5 |
* | | | RR | -43, 01, 00, 17 | 4,5 |
* | | | RRR | -043, 001, 123, 999, -999 | 4,5 |
* | | | RRRR | -0043, 0001, 2017, 9999, -9999 | 4,5 |
* | | | RRRRR | ... | 2,4,5 |
* | Extended year | 130 | u | -43, 1, 1900, 2017, 9999, -999 | 4 |
* | | | uu | -43, 01, 99, -99 | 4 |
* | | | uuu | -043, 001, 123, 999, -999 | 4 |
* | | | uuuu | -0043, 0001, 2017, 9999, -9999 | 4 |
* | | | uuuuu | ... | 2,4 |
* | Quarter (formatting) | 120 | Q | 1, 2, 3, 4 | |
* | | | Qo | 1st, 2nd, 3rd, 4th | 5 |
* | | | QQ | 01, 02, 03, 04 | |
* | | | QQQ | Q1, Q2, Q3, Q4 | |
* | | | QQQQ | 1st quarter, 2nd quarter, ... | 2 |
* | | | QQQQQ | 1, 2, 3, 4 | 4 |
* | Quarter (stand-alone) | 120 | q | 1, 2, 3, 4 | |
* | | | qo | 1st, 2nd, 3rd, 4th | 5 |
* | | | qq | 01, 02, 03, 04 | |
* | | | qqq | Q1, Q2, Q3, Q4 | |
* | | | qqqq | 1st quarter, 2nd quarter, ... | 2 |
* | | | qqqqq | 1, 2, 3, 4 | 3 |
* | Month (formatting) | 110 | M | 1, 2, ..., 12 | |
* | | | Mo | 1st, 2nd, ..., 12th | 5 |
* | | | MM | 01, 02, ..., 12 | |
* | | | MMM | Jan, Feb, ..., Dec | |
* | | | MMMM | January, February, ..., December | 2 |
* | | | MMMMM | J, F, ..., D | |
* | Month (stand-alone) | 110 | L | 1, 2, ..., 12 | |
* | | | Lo | 1st, 2nd, ..., 12th | 5 |
* | | | LL | 01, 02, ..., 12 | |
* | | | LLL | Jan, Feb, ..., Dec | |
* | | | LLLL | January, February, ..., December | 2 |
* | | | LLLLL | J, F, ..., D | |
* | Local week of year | 100 | w | 1, 2, ..., 53 | |
* | | | wo | 1st, 2nd, ..., 53th | 5 |
* | | | ww | 01, 02, ..., 53 | |
* | ISO week of year | 100 | I | 1, 2, ..., 53 | 5 |
* | | | Io | 1st, 2nd, ..., 53th | 5 |
* | | | II | 01, 02, ..., 53 | 5 |
* | Day of month | 90 | d | 1, 2, ..., 31 | |
* | | | do | 1st, 2nd, ..., 31st | 5 |
* | | | dd | 01, 02, ..., 31 | |
* | Day of year | 90 | D | 1, 2, ..., 365, 366 | 7 |
* | | | Do | 1st, 2nd, ..., 365th, 366th | 5 |
* | | | DD | 01, 02, ..., 365, 366 | 7 |
* | | | DDD | 001, 002, ..., 365, 366 | |
* | | | DDDD | ... | 2 |
* | Day of week (formatting) | 90 | E..EEE | Mon, Tue, Wed, ..., Sun | |
* | | | EEEE | Monday, Tuesday, ..., Sunday | 2 |
* | | | EEEEE | M, T, W, T, F, S, S | |
* | | | EEEEEE | Mo, Tu, We, Th, Fr, Sa, Su | |
* | ISO day of week (formatting) | 90 | i | 1, 2, 3, ..., 7 | 5 |
* | | | io | 1st, 2nd, ..., 7th | 5 |
* | | | ii | 01, 02, ..., 07 | 5 |
* | | | iii | Mon, Tue, Wed, ..., Sun | 5 |
* | | | iiii | Monday, Tuesday, ..., Sunday | 2,5 |
* | | | iiiii | M, T, W, T, F, S, S | 5 |
* | | | iiiiii | Mo, Tu, We, Th, Fr, Sa, Su | 5 |
* | Local day of week (formatting) | 90 | e | 2, 3, 4, ..., 1 | |
* | | | eo | 2nd, 3rd, ..., 1st | 5 |
* | | | ee | 02, 03, ..., 01 | |
* | | | eee | Mon, Tue, Wed, ..., Sun | |
* | | | eeee | Monday, Tuesday, ..., Sunday | 2 |
* | | | eeeee | M, T, W, T, F, S, S | |
* | | | eeeeee | Mo, Tu, We, Th, Fr, Sa, Su | |
* | Local day of week (stand-alone) | 90 | c | 2, 3, 4, ..., 1 | |
* | | | co | 2nd, 3rd, ..., 1st | 5 |
* | | | cc | 02, 03, ..., 01 | |
* | | | ccc | Mon, Tue, Wed, ..., Sun | |
* | | | cccc | Monday, Tuesday, ..., Sunday | 2 |
* | | | ccccc | M, T, W, T, F, S, S | |
* | | | cccccc | Mo, Tu, We, Th, Fr, Sa, Su | |
* | AM, PM | 80 | a..aaa | AM, PM | |
* | | | aaaa | a.m., p.m. | 2 |
* | | | aaaaa | a, p | |
* | AM, PM, noon, midnight | 80 | b..bbb | AM, PM, noon, midnight | |
* | | | bbbb | a.m., p.m., noon, midnight | 2 |
* | | | bbbbb | a, p, n, mi | |
* | Flexible day period | 80 | B..BBB | at night, in the morning, ... | |
* | | | BBBB | at night, in the morning, ... | 2 |
* | | | BBBBB | at night, in the morning, ... | |
* | Hour [1-12] | 70 | h | 1, 2, ..., 11, 12 | |
* | | | ho | 1st, 2nd, ..., 11th, 12th | 5 |
* | | | hh | 01, 02, ..., 11, 12 | |
* | Hour [0-23] | 70 | H | 0, 1, 2, ..., 23 | |
* | | | Ho | 0th, 1st, 2nd, ..., 23rd | 5 |
* | | | HH | 00, 01, 02, ..., 23 | |
* | Hour [0-11] | 70 | K | 1, 2, ..., 11, 0 | |
* | | | Ko | 1st, 2nd, ..., 11th, 0th | 5 |
* | | | KK | 01, 02, ..., 11, 00 | |
* | Hour [1-24] | 70 | k | 24, 1, 2, ..., 23 | |
* | | | ko | 24th, 1st, 2nd, ..., 23rd | 5 |
* | | | kk | 24, 01, 02, ..., 23 | |
* | Minute | 60 | m | 0, 1, ..., 59 | |
* | | | mo | 0th, 1st, ..., 59th | 5 |
* | | | mm | 00, 01, ..., 59 | |
* | Second | 50 | s | 0, 1, ..., 59 | |
* | | | so | 0th, 1st, ..., 59th | 5 |
* | | | ss | 00, 01, ..., 59 | |
* | Seconds timestamp | 40 | t | 512969520 | |
* | | | tt | ... | 2 |
* | Fraction of second | 30 | S | 0, 1, ..., 9 | |
* | | | SS | 00, 01, ..., 99 | |
* | | | SSS | 000, 001, ..., 999 | |
* | | | SSSS | ... | 2 |
* | Milliseconds timestamp | 20 | T | 512969520900 | |
* | | | TT | ... | 2 |
* | Timezone (ISO-8601 w/ Z) | 10 | X | -08, +0530, Z | |
* | | | XX | -0800, +0530, Z | |
* | | | XXX | -08:00, +05:30, Z | |
* | | | XXXX | -0800, +0530, Z, +123456 | 2 |
* | | | XXXXX | -08:00, +05:30, Z, +12:34:56 | |
* | Timezone (ISO-8601 w/o Z) | 10 | x | -08, +0530, +00 | |
* | | | xx | -0800, +0530, +0000 | |
* | | | xxx | -08:00, +05:30, +00:00 | 2 |
* | | | xxxx | -0800, +0530, +0000, +123456 | |
* | | | xxxxx | -08:00, +05:30, +00:00, +12:34:56 | |
* | Long localized date | NA | P | 05/29/1453 | 5,8 |
* | | | PP | May 29, 1453 | |
* | | | PPP | May 29th, 1453 | |
* | | | PPPP | Sunday, May 29th, 1453 | 2,5,8 |
* | Long localized time | NA | p | 12:00 AM | 5,8 |
* | | | pp | 12:00:00 AM | |
* | Combination of date and time | NA | Pp | 05/29/1453, 12:00 AM | |
* | | | PPpp | May 29, 1453, 12:00:00 AM | |
* | | | PPPpp | May 29th, 1453 at ... | |
* | | | PPPPpp | Sunday, May 29th, 1453 at ... | 2,5,8 |
* Notes:
* 1. "Formatting" units (e.g. formatting quarter) in the default en-US locale
* are the same as "stand-alone" units, but are different in some languages.
* "Formatting" units are declined according to the rules of the language
* in the context of a date. "Stand-alone" units are always nominative singular.
* In `format` function, they will produce different result:
*
* `format(new Date(2017, 10, 6), 'do LLLL', {locale: cs}) //=> '6. listopad'`
*
* `format(new Date(2017, 10, 6), 'do MMMM', {locale: cs}) //=> '6. listopadu'`
*
* `parse` will try to match both formatting and stand-alone units interchangably.
*
* 2. Any sequence of the identical letters is a pattern, unless it is escaped by
* the single quote characters (see below).
* If the sequence is longer than listed in table:
* - for numerical units (`yyyyyyyy`) `parse` will try to match a number
* as wide as the sequence
* - for text units (`MMMMMMMM`) `parse` will try to match the widest variation of the unit.
* These variations are marked with "2" in the last column of the table.
*
* 3. `QQQQQ` and `qqqqq` could be not strictly numerical in some locales.
* These tokens represent the shortest form of the quarter.
*
* 4. The main difference between `y` and `u` patterns are B.C. years:
*
* | Year | `y` | `u` |
* |------|-----|-----|
* | AC 1 | 1 | 1 |
* | BC 1 | 1 | 0 |
* | BC 2 | 2 | -1 |
*
* Also `yy` will try to guess the century of two digit year by proximity with `referenceDate`:
*
* `parse('50', 'yy', new Date(2018, 0, 1)) //=> Sat Jan 01 2050 00:00:00`
*
* `parse('75', 'yy', new Date(2018, 0, 1)) //=> Wed Jan 01 1975 00:00:00`
*
* while `uu` will just assign the year as is:
*
* `parse('50', 'uu', new Date(2018, 0, 1)) //=> Sat Jan 01 0050 00:00:00`
*
* `parse('75', 'uu', new Date(2018, 0, 1)) //=> Tue Jan 01 0075 00:00:00`
*
* The same difference is true for local and ISO week-numbering years (`Y` and `R`),
* except local week-numbering years are dependent on `options.weekStartsOn`
* and `options.firstWeekContainsDate` (compare [setISOWeekYear](https://date-fns.org/docs/setISOWeekYear)
* and [setWeekYear](https://date-fns.org/docs/setWeekYear)).
*
* 5. These patterns are not in the Unicode Technical Standard #35:
* - `i`: ISO day of week
* - `I`: ISO week of year
* - `R`: ISO week-numbering year
* - `o`: ordinal number modifier
* - `P`: long localized date
* - `p`: long localized time
*
* 6. `YY` and `YYYY` tokens represent week-numbering years but they are often confused with years.
* You should enable `options.useAdditionalWeekYearTokens` to use them. See: https://github.com/date-fns/date-fns/blob/master/docs/unicodeTokens.md
*
* 7. `D` and `DD` tokens represent days of the year but they are ofthen confused with days of the month.
* You should enable `options.useAdditionalDayOfYearTokens` to use them. See: https://github.com/date-fns/date-fns/blob/master/docs/unicodeTokens.md
*
* 8. `P+` tokens do not have a defined priority since they are merely aliases to other tokens based
* on the given locale.
*
* using `en-US` locale: `P` => `MM/dd/yyyy`
* using `en-US` locale: `p` => `hh:mm a`
* using `pt-BR` locale: `P` => `dd/MM/yyyy`
* using `pt-BR` locale: `p` => `HH:mm`
*
* Values will be assigned to the date in the descending order of its unit's priority.
* Units of an equal priority overwrite each other in the order of appearance.
*
* If no values of higher priority are parsed (e.g. when parsing string 'January 1st' without a year),
* the values will be taken from 3rd argument `referenceDate` which works as a context of parsing.
*
* `referenceDate` must be passed for correct work of the function.
* If you're not sure which `referenceDate` to supply, create a new instance of Date:
* `parse('02/11/2014', 'MM/dd/yyyy', new Date())`
* In this case parsing will be done in the context of the current date.
* If `referenceDate` is `Invalid Date` or a value not convertible to valid `Date`,
* then `Invalid Date` will be returned.
*
* The result may vary by locale.
*
* If `formatString` matches with `dateString` but does not provides tokens, `referenceDate` will be returned.
*
* If parsing failed, `Invalid Date` will be returned.
* Invalid Date is a Date, whose time value is NaN.
* Time value of Date: http://es5.github.io/#x15.9.1.1
*
* @typeParam DateType - The `Date` type, the function operates on. Gets inferred from passed arguments. Allows to use extensions like [`UTCDate`](https://github.com/date-fns/utc).
*
* @param dateStr - The string to parse
* @param formatStr - The string of tokens
* @param referenceDate - defines values missing from the parsed dateString
* @param options - An object with options.
* see: https://github.com/date-fns/date-fns/blob/master/docs/unicodeTokens.md
* see: https://github.com/date-fns/date-fns/blob/master/docs/unicodeTokens.md
*
* @returns The parsed date
*
* @throws `options.locale` must contain `match` property
* @throws use `yyyy` instead of `YYYY` for formatting years using [format provided] to the input [input provided]; see: https://github.com/date-fns/date-fns/blob/master/docs/unicodeTokens.md
* @throws use `yy` instead of `YY` for formatting years using [format provided] to the input [input provided]; see: https://github.com/date-fns/date-fns/blob/master/docs/unicodeTokens.md
* @throws use `d` instead of `D` for formatting days of the month using [format provided] to the input [input provided]; see: https://github.com/date-fns/date-fns/blob/master/docs/unicodeTokens.md
* @throws use `dd` instead of `DD` for formatting days of the month using [format provided] to the input [input provided]; see: https://github.com/date-fns/date-fns/blob/master/docs/unicodeTokens.md
* @throws format string contains an unescaped latin alphabet character
*
* @example
* // Parse 11 February 2014 from middle-endian format:
* var result = parse('02/11/2014', 'MM/dd/yyyy', new Date())
* //=> Tue Feb 11 2014 00:00:00
*
* @example
* // Parse 28th of February in Esperanto locale in the context of 2010 year:
* import eo from 'date-fns/locale/eo'
* var result = parse('28-a de februaro', "do 'de' MMMM", new Date(2010, 0, 1), {
* locale: eo
* })
* //=> Sun Feb 28 2010 00:00:00
*/
export function parse(dateStr, formatStr, referenceDate, options) {
const defaultOptions = getDefaultOptions();
const locale = options?.locale ?? defaultOptions.locale ?? defaultLocale;
const firstWeekContainsDate =
options?.firstWeekContainsDate ??
options?.locale?.options?.firstWeekContainsDate ??
defaultOptions.firstWeekContainsDate ??
defaultOptions.locale?.options?.firstWeekContainsDate ??
1;
const weekStartsOn =
options?.weekStartsOn ??
options?.locale?.options?.weekStartsOn ??
defaultOptions.weekStartsOn ??
defaultOptions.locale?.options?.weekStartsOn ??
0;
if (formatStr === "") {
if (dateStr === "") {
return toDate(referenceDate);
} else {
return constructFrom(referenceDate, NaN);
}
}
const subFnOptions = {
firstWeekContainsDate,
weekStartsOn,
locale,
};
// If timezone isn't specified, it will be set to the system timezone
const setters = [new DateToSystemTimezoneSetter()];
const tokens = formatStr
.match(longFormattingTokensRegExp)
.map((substring) => {
const firstCharacter = substring[0];
if (firstCharacter in longFormatters) {
const longFormatter = longFormatters[firstCharacter];
return longFormatter(substring, locale.formatLong);
}
return substring;
})
.join("")
.match(formattingTokensRegExp);
const usedTokens = [];
for (let token of tokens) {
if (
!options?.useAdditionalWeekYearTokens &&
isProtectedWeekYearToken(token)
) {
warnOrThrowProtectedError(token, formatStr, dateStr);
}
if (
!options?.useAdditionalDayOfYearTokens &&
isProtectedDayOfYearToken(token)
) {
warnOrThrowProtectedError(token, formatStr, dateStr);
}
const firstCharacter = token[0];
const parser = parsers[firstCharacter];
if (parser) {
const { incompatibleTokens } = parser;
if (Array.isArray(incompatibleTokens)) {
const incompatibleToken = usedTokens.find(
(usedToken) =>
incompatibleTokens.includes(usedToken.token) ||
usedToken.token === firstCharacter,
);
if (incompatibleToken) {
throw new RangeError(
`The format string mustn't contain \`${incompatibleToken.fullToken}\` and \`${token}\` at the same time`,
);
}
} else if (parser.incompatibleTokens === "*" && usedTokens.length > 0) {
throw new RangeError(
`The format string mustn't contain \`${token}\` and any other token at the same time`,
);
}
usedTokens.push({ token: firstCharacter, fullToken: token });
const parseResult = parser.run(
dateStr,
token,
locale.match,
subFnOptions,
);
if (!parseResult) {
return constructFrom(referenceDate, NaN);
}
setters.push(parseResult.setter);
dateStr = parseResult.rest;
} else {
if (firstCharacter.match(unescapedLatinCharacterRegExp)) {
throw new RangeError(
"Format string contains an unescaped latin alphabet character `" +
firstCharacter +
"`",
);
}
// Replace two single quote characters with one single quote character
if (token === "''") {
token = "'";
} else if (firstCharacter === "'") {
token = cleanEscapedString(token);
}
// Cut token from string, or, if string doesn't match the token, return Invalid Date
if (dateStr.indexOf(token) === 0) {
dateStr = dateStr.slice(token.length);
} else {
return constructFrom(referenceDate, NaN);
}
}
}
// Check if the remaining input contains something other than whitespace
if (dateStr.length > 0 && notWhitespaceRegExp.test(dateStr)) {
return constructFrom(referenceDate, NaN);
}
const uniquePrioritySetters = setters
.map((setter) => setter.priority)
.sort((a, b) => b - a)
.filter((priority, index, array) => array.indexOf(priority) === index)
.map((priority) =>
setters
.filter((setter) => setter.priority === priority)
.sort((a, b) => b.subPriority - a.subPriority),
)
.map((setterArray) => setterArray[0]);
let date = toDate(referenceDate);
if (isNaN(date.getTime())) {
return constructFrom(referenceDate, NaN);
}
const flags = {};
for (const setter of uniquePrioritySetters) {
if (!setter.validate(date, subFnOptions)) {
return constructFrom(referenceDate, NaN);
}
const result = setter.set(date, flags, subFnOptions);
// Result is tuple (date, flags)
if (Array.isArray(result)) {
date = result[0];
Object.assign(flags, result[1]);
// Result is date
} else {
date = result;
}
}
return constructFrom(referenceDate, date);
}
function cleanEscapedString(input) {
return input.match(escapedStringRegExp)[1].replace(doubleQuoteRegExp, "'");
}
// Fallback for modularized imports:
export default parse;