Improve extractNumberFromString() for international number formats (#2843)

* Set up test

* Add support for international formatted numbers

* Expose locale in scraper configuration

* Update changelog

CHANGELOG.md

@@ -10,6 +10,8 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 ### Added
 
 - Broken down the performance into asset and currency on the analysis page (experimental)
+- Added support for international formatted numbers in the scraper configuration
+- Added the attribute `locale` to the scraper configuration to parse the number
 
 ### Changed
 

apps/api/src/app/info/info.service.ts

@@ -195,11 +195,11 @@ export class InfoService {
 
       const $ = cheerio.load(body);
 
-      return extractNumberFromString(
-        $(
+      return extractNumberFromString({
+        value: $(
           `a[href="/ghostfolio/ghostfolio/graphs/contributors"] .Counter`
         ).text()
-      );
+      });
     } catch (error) {
       Logger.error(error, 'InfoService - GitHub');
 

apps/api/src/services/data-provider/manual/manual.service.ts

@@ -236,6 +236,7 @@ export class ManualService implements DataProviderInterface {
         abortController.abort();
       }, this.configurationService.get('REQUEST_TIMEOUT'));
 
+      let locale = scraperConfiguration.locale;
       const { body, headers } = await got(scraperConfiguration.url, {
         headers: scraperConfiguration.headers as Headers,
         // @ts-ignore
@@ -248,13 +249,20 @@ export class ManualService implements DataProviderInterface {
           jsonpath.query(data, scraperConfiguration.selector)[0]
         );
 
-        return extractNumberFromString(value);
+        return extractNumberFromString({ locale, value });
       } else {
         const $ = cheerio.load(body);
 
-        return extractNumberFromString(
-          $(scraperConfiguration.selector).first().text()
-        );
+        if (!locale) {
+          try {
+            locale = $('html').attr('lang');
+          } catch {}
+        }
+
+        return extractNumberFromString({
+          locale,
+          value: $(scraperConfiguration.selector).first().text()
+        });
       }
     } catch (error) {
       throw error;

apps/api/src/services/symbol-profile/symbol-profile.service.ts

@@ -202,6 +202,7 @@ export class SymbolProfileService {
         defaultMarketPrice: scraperConfiguration.defaultMarketPrice as number,
         headers:
           scraperConfiguration.headers as ScraperConfiguration['headers'],
+        locale: scraperConfiguration.locale as string,
         selector: scraperConfiguration.selector as string,
         url: scraperConfiguration.url as string
       };

libs/common/src/lib/helper.spec.ts

@@ -0,0 +1,39 @@
+import { extractNumberFromString } from '@ghostfolio/common/helper';
+
+describe('Helper', () => {
+  describe('Extract number from string', () => {
+    it('Get decimal number', async () => {
+      expect(extractNumberFromString({ value: '999.99' })).toEqual(999.99);
+    });
+
+    it('Get decimal number (with spaces)', async () => {
+      expect(extractNumberFromString({ value: ' 999.99 ' })).toEqual(999.99);
+    });
+
+    it('Get decimal number (with currency)', async () => {
+      expect(extractNumberFromString({ value: '999.99 CHF' })).toEqual(999.99);
+    });
+
+    it('Get decimal number (comma notation)', async () => {
+      expect(
+        extractNumberFromString({ locale: 'de-DE', value: '999,99' })
+      ).toEqual(999.99);
+    });
+
+    it('Get decimal number with group (dot notation)', async () => {
+      expect(
+        extractNumberFromString({ locale: 'de-CH', value: '99’999.99' })
+      ).toEqual(99999.99);
+    });
+
+    it('Get decimal number with group (comma notation)', async () => {
+      expect(
+        extractNumberFromString({ locale: 'de-DE', value: '99.999,99' })
+      ).toEqual(99999.99);
+    });
+
+    it('Not a number', async () => {
+      expect(extractNumberFromString({ value: 'X' })).toEqual(NaN);
+    });
+  });
+});

libs/common/src/lib/helper.ts

@@ -1,4 +1,5 @@
 import * as currencies from '@dinero.js/currencies';
+import { NumberParser } from '@internationalized/number';
 import { DataSource, MarketData } from '@prisma/client';
 import Big from 'big.js';
 import {
@@ -20,8 +21,6 @@ export const DATE_FORMAT = 'yyyy-MM-dd';
 export const DATE_FORMAT_MONTHLY = 'MMMM yyyy';
 export const DATE_FORMAT_YEARLY = 'yyyy';
 
-const NUMERIC_REGEXP = /[-]{0,1}[\d]*[.,]{0,1}[\d]+/g;
-
 export function calculateBenchmarkTrend({
   days,
   historicalData
@@ -120,10 +119,20 @@ export function encodeDataSource(aDataSource: DataSource) {
   return undefined;
 }
 
-export function extractNumberFromString(aString: string): number {
+export function extractNumberFromString({
+  locale = 'en-US',
+  value
+}: {
+  locale?: string;
+  value: string;
+}): number {
   try {
-    const [numberString] = aString.match(NUMERIC_REGEXP);
-    return parseFloat(numberString.trim());
+    // Remove non-numeric characters (excluding international formatting characters)
+    const numericValue = value.replace(/[^\d.,'’\s]/g, '');
+
+    let parser = new NumberParser(locale);
+
+    return parser.parse(numericValue);
   } catch {
     return undefined;
   }

libs/common/src/lib/interfaces/scraper-configuration.interface.ts

@@ -1,6 +1,7 @@
 export interface ScraperConfiguration {
   defaultMarketPrice?: number;
   headers?: { [key: string]: string };
+  locale?: string;
   selector: string;
   url: string;
 }

package.json

@@ -44,7 +44,9 @@
     "start:production": "yarn database:migrate && yarn database:seed && node main",
     "start:server": "nx run api:serve --watch",
     "start:storybook": "nx run ui:storybook",
-    "test": "npx dotenv-cli -e .env.example -- nx test",
+    "test": "yarn test:api && yarn test:common",
+    "test:api": "npx dotenv-cli -e .env.example -- nx test api",
+    "test:common": "npx dotenv-cli -e .env.example -- nx test common",
     "test:single": "nx run api:test --test-file portfolio-calculator-novn-buy-and-sell.spec.ts",
     "ts-node": "ts-node",
     "update": "nx migrate latest",
@@ -71,6 +73,7 @@
     "@dfinity/identity": "0.15.7",
     "@dfinity/principal": "0.15.7",
     "@dinero.js/currencies": "2.0.0-alpha.8",
+    "@internationalized/number": "3.5.0",
     "@nestjs/bull": "10.0.1",
     "@nestjs/cache-manager": "2.1.0",
     "@nestjs/common": "10.1.3",

yarn.lock

@@ -3063,6 +3063,13 @@
   resolved "https://registry.yarnpkg.com/@humanwhocodes/object-schema/-/object-schema-2.0.1.tgz#e5211452df060fa8522b55c7b3c0c4d1981cb044"
   integrity sha512-dvuCeX5fC9dXgJn9t+X5atfmgQAzUOWqS1254Gh0m6i8wKd10ebXkfNKiRK+1GWi/yTvvLDHpoxLr0xxxeslWw==
 
+"@internationalized/number@3.5.0":
+  version "3.5.0"
+  resolved "https://registry.yarnpkg.com/@internationalized/number/-/number-3.5.0.tgz#9de6018424b441a6545f209afa286ad7df4a2906"
+  integrity sha512-ZY1BW8HT9WKYvaubbuqXbbDdHhOUMfE2zHHFJeTppid0S+pc8HtdIxFxaYMsGjCb4UsF+MEJ4n2TfU7iHnUK8w==
+  dependencies:
+    "@swc/helpers" "^0.5.0"
+
 "@ioredis/commands@^1.1.1":
   version "1.2.0"
   resolved "https://registry.yarnpkg.com/@ioredis/commands/-/commands-1.2.0.tgz#6d61b3097470af1fdbbe622795b8921d42018e11"
@@ -6201,6 +6208,13 @@
   resolved "https://registry.yarnpkg.com/@swc/counter/-/counter-0.1.2.tgz#bf06d0770e47c6f1102270b744e17b934586985e"
   integrity sha512-9F4ys4C74eSTEUNndnER3VJ15oru2NumfQxS8geE+f3eB5xvfxpWyqE5XlVnxb/R14uoXi6SLbBwwiDSkv+XEw==
 
+"@swc/helpers@^0.5.0":
+  version "0.5.3"
+  resolved "https://registry.yarnpkg.com/@swc/helpers/-/helpers-0.5.3.tgz#98c6da1e196f5f08f977658b80d6bd941b5f294f"
+  integrity sha512-FaruWX6KdudYloq1AHD/4nU+UsMTdNE8CKyrseXWEcgjDAbvkwJg2QGPAnfIJLIWsjZOSPLOAykK6fuYp4vp4A==
+  dependencies:
+    tslib "^2.4.0"
+
 "@swc/types@^0.1.5":
   version "0.1.5"
   resolved "https://registry.yarnpkg.com/@swc/types/-/types-0.1.5.tgz#043b731d4f56a79b4897a3de1af35e75d56bc63a"