🔀 Merge branch 'release/2.0.2'

CHANGELOG.md

@@ -1,5 +1,11 @@
 # Changelog
 
+## [2.0.2] - 2019-06-09
+
+### Added
+
+- Ability to parse more date formats, including Finnish, that look something like this: `31.5.2019 klo 16.58 - <author>: <message>`
+
 ## [2.0.1] - 2019-05-31
 
 ### Changed

README.md

@@ -55,7 +55,7 @@ You can also use the [jsDelivr CDN](https://www.jsdelivr.com/package/npm/whatsap
 ```html
 <script src="https://cdn.jsdelivr.net/npm/whatsapp-chat-parser/dist/whatsapp-chat-parser.min.js"></script>
 <!-- Or use a specific version -->
-<script src="https://cdn.jsdelivr.net/npm/whatsapp-chat-parser@2.0.1/dist/whatsapp-chat-parser.min.js"></script>
+<script src="https://cdn.jsdelivr.net/npm/whatsapp-chat-parser@2.0.2/dist/whatsapp-chat-parser.min.js"></script>
 ```
 
 &nbsp;

package.json

@@ -1,6 +1,6 @@
 {
   "name": "whatsapp-chat-parser",
-  "version": "2.0.1",
+  "version": "2.0.2",
   "description": "A package to parse WhatsApp chats",
   "main": "dist/whatsapp-chat-parser.js",
   "files": [
@@ -9,9 +9,9 @@
   "scripts": {
     "build": "rm -rf dist/ && rollup -c",
     "pretest": "npm run build",
-    "test": "jest --passWithNoTests",
-    "test:watch": "jest --watchAll",
-    "test:coverage": "jest --coverage",
+    "test": "TZ=UTC jest --passWithNoTests",
+    "test:watch": "TZ=UTC jest --watchAll",
+    "test:coverage": "TZ=UTC jest --coverage",
     "lint": "eslint src/**/*.js",
     "format": "pretty-quick",
     "format:staged": "pretty-quick --staged",
@@ -49,11 +49,11 @@
     "eslint-config-prettier": "4.3.0",
     "eslint-plugin-import": "2.17.3",
     "eslint-plugin-prettier": "3.1.0",
-    "husky": "2.3.0",
+    "husky": "2.4.0",
     "jest": "24.8.0",
-    "prettier": "1.17.1",
+    "prettier": "1.18.2",
     "pretty-quick": "1.11.0",
-    "rollup": "1.12.5",
+    "rollup": "1.14.5",
     "rollup-plugin-babel": "4.3.2",
     "rollup-plugin-commonjs": "10.0.0",
     "rollup-plugin-uglify": "6.0.2"

src/parser.js

@@ -1,13 +1,15 @@
 const { daysBeforeMonths, normalizeDate } = require('./date.js');
 const {
+  regexSplitTime,
   convertTime12to24,
   normalizeAMPM,
   normalizeTime,
 } = require('./time.js');
 
-const regexParser = /\[?(\d{1,2}[-/.]\d{1,2}[-/.]\d{2,4}),? (\d{1,2}[.:]\d{1,2}(?:[.:]\d{1,2})?)(?: ([ap]\.?m\.?))?\]?(?: -|:)? (.+?): ([^]*)/i;
-const regexParserSystem = /\[?(\d{1,2}[-/.]\d{1,2}[-/.]\d{2,4}),? (\d{1,2}[.:]\d{1,2}(?:[.:]\d{1,2})?)(?: ([ap]\.?m\.?))?\]?(?: -|:)? ([^]+)/i;
-const regexStartsWithDateTime = /\[?(\d{1,2}[-/.]\d{1,2}[-/.]\d{2,4}),? (\d{1,2}[.:]\d{1,2}(?:[.:]\d{1,2})?)(?: ([ap]\.?m\.?))?\]?/i;
+const regexParser = /\[?(\d{1,2}[-/.] ?\d{1,2}[-/.] ?\d{2,4})[,.]? \D*?(\d{1,2}[.:]\d{1,2}(?:[.:]\d{1,2})?)(?: ([ap]\.?m\.?))?\]?(?: -|:)? (.+?): ([^]*)/i;
+const regexParserSystem = /\[?(\d{1,2}[-/.] ?\d{1,2}[-/.] ?\d{2,4})[,.]? \D*?(\d{1,2}[.:]\d{1,2}(?:[.:]\d{1,2})?)(?: ([ap]\.?m\.?))?\]?(?: -|:)? ([^]+)/i;
+const regexStartsWithDateTime = /\[?(\d{1,2}[-/.] ?\d{1,2}[-/.] ?\d{2,4})[,.]? \D*?(\d{1,2}[.:]\d{1,2}(?:[.:]\d{1,2})?)(?: ([ap]\.?m\.?))?\]?/i;
+const regexSplitDate = /[-/.] ?/;
 
 /**
  * Given an array of lines, detects the lines that are part of a previous
@@ -74,7 +76,7 @@ function parseMessages(messages, options = { daysFirst: undefined }) {
   if (typeof daysFirst !== 'boolean') {
     const numericDates = Array.from(
       new Set(parsed.map(({ date }) => date)),
-      date => date.split(/[-/.]/).map(Number),
+      date => date.split(regexSplitDate).map(Number),
     );
 
     daysFirst = daysBeforeMonths(numericDates);
@@ -87,16 +89,16 @@ function parseMessages(messages, options = { daysFirst: undefined }) {
     let year;
 
     if (daysFirst === false) {
-      [month, day, year] = date.split(/[-/.]/);
+      [month, day, year] = date.split(regexSplitDate);
     } else {
-      [day, month, year] = date.split(/[-/.]/);
+      [day, month, year] = date.split(regexSplitDate);
     }
 
     [year, month, day] = normalizeDate(year, month, day);
 
     const [hours, minutes, seconds] = normalizeTime(
       ampm ? convertTime12to24(time, normalizeAMPM(ampm)) : time,
-    ).split(/[:.]/);
+    ).split(regexSplitTime);
 
     return {
       date: new Date(year, month - 1, day, hours, minutes, seconds),

src/parser.test.js

@@ -131,22 +131,28 @@ describe('parser.js', () => {
       const format1 = [{ system: false, msg: '3/6/18, 1:55 p.m. - a: m' }];
       const format2 = [{ system: false, msg: '03-06-2018, 01.55 PM - a: m' }];
       const format3 = [{ system: false, msg: '13.06.18 21.25.15: a: m' }];
-      const format4 = [{ system: false, msg: '[03.13.18 21:25:15] a: m' }];
+      const format4 = [{ system: false, msg: '[06.13.18 21:25:15] a: m' }];
+      const format5 = [{ system: false, msg: '13.6.2018 klo 21.25.15 - a: m' }];
+      const format6 = [{ system: false, msg: '13. 6. 2018. 21:25:15 a: m' }];
+
       const parsed1 = parseMessages(format1);
       const parsed2 = parseMessages(format2);
       const parsed3 = parseMessages(format3);
       const parsed4 = parseMessages(format4);
+      const parsed5 = parseMessages(format5);
+      const parsed6 = parseMessages(format6);
+
+      const expected1 = '2018-06-03T13:55:00.000Z';
+      const expected2 = '2018-06-13T21:25:15.000Z';
 
       describe('the date', () => {
         it('should be parsed correctly in various formats', () => {
-          /**
-           * Checking for the year should be enough to know there were no errors
-           * in parsing a specific format
-           */
-          expect(parsed1[0].date.getFullYear()).toBe(2018);
-          expect(parsed2[0].date.getFullYear()).toBe(2018);
-          expect(parsed3[0].date.getFullYear()).toBe(2018);
-          expect(parsed4[0].date.getFullYear()).toBe(2018);
+          expect(parsed1[0].date.toISOString()).toBe(expected1);
+          expect(parsed2[0].date.toISOString()).toBe(expected1);
+          expect(parsed3[0].date.toISOString()).toBe(expected2);
+          expect(parsed4[0].date.toISOString()).toBe(expected2);
+          expect(parsed5[0].date.toISOString()).toBe(expected2);
+          expect(parsed6[0].date.toISOString()).toBe(expected2);
         });
       });
     });

src/time.js

@@ -1,10 +1,12 @@
+const regexSplitTime = /[:.]/;
+
 /**
  * Converts time from 12 hour format to 24 hour format
  * From: https://stackoverflow.com/a/40197728/5303634
  */
 function convertTime12to24(time, ampm) {
   // eslint-disable-next-line prefer-const
-  let [hours, minutes, seconds] = time.split(/[:.]/);
+  let [hours, minutes, seconds] = time.split(regexSplitTime);
 
   if (hours === '12') {
     hours = '00';
@@ -21,7 +23,7 @@ function convertTime12to24(time, ampm) {
  * Normalizes a time string to have the following format: hh:mm:ss
  */
 function normalizeTime(time) {
-  const [hours, minutes, seconds] = time.split(/[:.]/);
+  const [hours, minutes, seconds] = time.split(regexSplitTime);
 
   return `${hours.length === 1 ? `0${hours}` : hours}:${minutes}:${seconds ||
     '00'}`;
@@ -34,4 +36,9 @@ function normalizeAMPM(ampm) {
   return ampm.replace(/[^apm]/gi, '').toUpperCase();
 }
 
-module.exports = { convertTime12to24, normalizeTime, normalizeAMPM };
+module.exports = {
+  regexSplitTime,
+  convertTime12to24,
+  normalizeTime,
+  normalizeAMPM,
+};