Skip to content

Latest commit

 

History

History
136 lines (93 loc) · 5.94 KB

README-EN.md

File metadata and controls

136 lines (93 loc) · 5.94 KB
Raw

@alordash/date-parser

$ npm i @alordash/date-parser

Links

Description

Extracts dates of any event from human spoken message with seconds precision.
Can return event description without date-related words.
It uses my realization of Damerau-Levenshtein algorithm to properly parse words even if they are written with mistakes.
Supports Russian and English languages.
Easy to use, this library allows you to determine dates with absolute precision using a variety of parameters.

Documentaion

TimeList class:

Each property is Number that represents value of specific time type.
If value is undefined, then source string did not contain definition of that time type.

Properties:

  • TimeList.dates {Number} — Date.setDate(TimeList.dates);
  • TimeList.hours {Number} — Date.setHours(TimeList.hours);
  • TimeList.minutes {Number} — Date.setMinutes(TimeList.minutes);
  • TimeList.months {Number} — Date.setMonth(TimeList.months);
  • TimeList.seconds {Number} — Date.setSeconds(TimeList.seconds);
  • TimeList.years {Number} — Date.setFullYear(TimeList.years);

ParsedDate class:

Every ParsedDate object represents particular date and event retrieved from source string.

Properties:

  • ParsedDate.confidence {Number} — confidence level of this ParsedDate.
  • ParsedDate.contexts {Array.<Number>} — technical property.
  • ParsedDate.target_date {TimeList} — contains all information about target date of event.
  • ParsedDate.period_time {TimeList} — contains all information about periodicity of event.
  • ParsedDate.max_date {TimeList} — contains all information about maximum date of event.
  • ParsedDate.string {String} — event description without date-related words.
  • ParsedDate.isOffset {Boolean} — is this date determined with offset (using word "in").
  • ParsedDate.isFixed {Boolean} — does contain dates with unambiguous definition (e.g. "9 a.m.").

Function parseDate(string, errorLimit, minimumPrevalence): {Array.<ParsedDates>}

Arguments

  1. string {String} — source string.
  2. errorLimit {Number} — From 0.0 to 1.0, the less — the less results. Used for recognizing words with mistakes.
  3. minimumPrevalence {Number} — From 0 to 100, the less — the more results. Used to filter rare date cases.

Usage

const { parseDate } = require('@alordash/date-parser');

let string = 'Come home at 8:30 p.m. and cook dinner. Buy milk and wash car on monday.';
let result = parseDate(string);

console.log('time 1 :>> ', JSON.stringify(result[0].target_date));
//=> time 1 :>>  {"minutes":30,"hours":20}

console.log('event 1 :>> ', result[0].string);
//=> event 1 :>> Come home and cook dinner

console.log('time 2 :>> ', JSON.stringify(result[1].target_date));
//=> time 2 :>>  {"dates":31} //30.08.20 atm., 31.08.20 is Monday

console.log('event 2 :>> ', result[1].string);
//=> event 2 :>> Buy milk and wash car

//————————————————————————————————————————————————————————————————————————————————————————————————————————————————————————————————
string = 'Visit doctor from 9 a.m. to 11 p.m. on next Saturday and go to shop at 7 p.m.';
result = parseDate(string);

console.log('maximum date 1 :>> ', JSON.stringify(result[0].max_date));
//=> maximum date 1 :>>  {"dates":36,"hours":23} //30.08.20 atm., next saturday is on 05.09.20, which is technically 36.08.20

console.log('target date 1 :>> ', JSON.stringify(result[0].target_date));
//=> target date 1 :>>  {"hours":9}

console.log('event 1 :>> ', result[0].string);
//=> event 1 :>> Visit doctor

console.log('target date 2 :>> ', JSON.stringify(result[1].target_date));
//=> target date 2 :>>  {"hours":19}

console.log('event 2 :>> ', result[1].string);
//=> event 2 :>> go to shop

//————————————————————————————————————————————————————————————————————————————————————————————————————————————————————————————————
string = 'On 19 of September From 9:00 to 20:00 Get up from the computer every 15 minutes and do a warm-up';
result = parseDate(string);

console.log('target date :>> ', JSON.stringify(result[0].target_date));
//=> target date :>>  {"dates":19,"hours":9,"minutes":0,"months":9}

console.log('period time :>> ', JSON.stringify(result[0].period_time));
//=> period time :>>  {"minutes":15}

console.log('maximum date :>> ', JSON.stringify(result[0].maximum_date));
//=> maximum date :>>  {"hours":20,"minutes":0}

console.log('event :>> ', JSON.stringify(result[0].string));
//=> event :>> Get up from the computer and do a warm-up

Features

ParsedDate.toString(): {String}

Returns event description.

ParsedDate.valueOf(): {{target_date: Date, period_time: Date, max_date: Date}}

Composes all found time types to dates and returns

{
    target_date: Date,  
    period_time: Date,  
    max_date: Date  
}

Using current time values (new Date()) for not found target_dates and max_dates time types.
Using null time values (new Date(0)) for not found period_time time types.

Custom settings

You can change csv files that are used for recognition via functions SetExpressionsDirectoryPath(directoryPath: string) and SetSeparatorsDirectoryPath(directoryPath: string).