test(lyra): removes skips to stemming tests

oramasearch · Jul 19, 2022 · 41069fa · 41069fa
1 parent 87a634f
commit 41069fa
Show file tree

Hide file tree

Showing 5 changed files with 70 additions and 20 deletions.
diff --git a/packages/lyra/package.json b/packages/lyra/package.json
@@ -22,7 +22,7 @@
   "dependencies": {
     "fastq": "^1.13.0",
     "nanoid": "3.3.4",
-    "natural": "^5.2.2"
+    "natural": "^5.2.3"
   },
   "devDependencies": {
     "@types/jest": "^27.5.0",

diff --git a/packages/lyra/src/stemmer.ts b/packages/lyra/src/stemmer.ts
@@ -1,4 +1,15 @@
-import natural, { Stemmer } from "natural";
+import {
+  PorterStemmer,
+  PorterStemmerNl,
+  PorterStemmerFr,
+  PorterStemmerEs,
+  PorterStemmerIt,
+  PorterStemmerPt,
+  PorterStemmerSv,
+  PorterStemmerRu,
+  PorterStemmerNo,
+  Stemmer,
+} from "natural";
 
 export type Language = typeof SUPPORTED_LANGUAGES[number];
 
@@ -19,34 +30,34 @@ export function stemArray(input: string[], language: Language): string[] {
 
   switch (language) {
     case "dutch":
-      stemmer = natural.PorterStemmerNl;
+      stemmer = PorterStemmerNl;
       break;
     case "english":
-      stemmer = natural.PorterStemmer;
+      stemmer = PorterStemmer;
       break;
     case "french":
-      stemmer = natural.PorterStemmerFr;
+      stemmer = PorterStemmerFr;
       break;
     case "italian":
-      stemmer = natural.PorterStemmerIt;
+      stemmer = PorterStemmerIt;
       break;
     case "norwegian":
-      stemmer = natural.PorterStemmerNo;
+      stemmer = PorterStemmerNo;
       break;
     case "portugese":
-      stemmer = natural.PorterStemmerPt;
+      stemmer = PorterStemmerPt;
       break;
     case "russian":
-      stemmer = natural.PorterStemmerRu;
+      stemmer = PorterStemmerRu;
       break;
     case "spanish":
-      stemmer = natural.PorterStemmerEs;
+      stemmer = PorterStemmerEs;
       break;
     case "swedish":
-      stemmer = natural.PorterStemmerSv;
+      stemmer = PorterStemmerSv;
       break;
     default:
-      stemmer = natural.PorterStemmer;
+      stemmer = PorterStemmer;
       break;
   }
 

diff --git a/packages/lyra/tests/__snapshots__/tokenizer.test.ts.snap b/packages/lyra/tests/__snapshots__/tokenizer.test.ts.snap
@@ -1,5 +1,23 @@
 // Jest Snapshot v1, https://goo.gl/fbAQLP
 
+exports[`Should tokenize and stem correctly in dutch 1`] = `
+Set {
+  "de",
+  "klein",
+  "koei",
+}
+`;
+
+exports[`Should tokenize and stem correctly in dutch 2`] = `
+Set {
+  "ik",
+  "heb",
+  "wat",
+  "taart",
+  "gemaakt",
+}
+`;
+
 exports[`Should tokenize and stem correctly in english 1`] = `
 Set {
   "the",
@@ -132,6 +150,27 @@ Set {
 }
 `;
 
+exports[`Should tokenize and stem correctly in spanish 1`] = `
+Set {
+  "cocin",
+  "unos",
+  "pastel",
+}
+`;
+
+exports[`Should tokenize and stem correctly in spanish 2`] = `
+Set {
+  "dorm",
+  "es",
+  "algo",
+  "dificil",
+  "cuand",
+  "las",
+  "prueb",
+  "fall",
+}
+`;
+
 exports[`Should tokenize and stem correctly in swedish 1`] = `
 Set {
   "jag",

diff --git a/packages/lyra/tests/tokenizer.test.ts b/packages/lyra/tests/tokenizer.test.ts
@@ -78,7 +78,7 @@ it("Should tokenize and stem correctly in swedish", () => {
   expect(O2).toMatchSnapshot();
 });
 
-it.skip("Should tokenize and stem correctly in spanish", () => {
+it("Should tokenize and stem correctly in spanish", () => {
   const I1 = "cociné unos pasteles";
   const I2 = "dormir es algo dificil cuando las pruebas fallan";
 
@@ -89,7 +89,7 @@ it.skip("Should tokenize and stem correctly in spanish", () => {
   expect(O2).toMatchSnapshot();
 });
 
-it.skip("Should tokenize and stem correctly in dutch", () => {
+it("Should tokenize and stem correctly in dutch", () => {
   const I1 = "de kleine koeien";
   const I2 = "Ik heb wat taarten gemaakt";
 

diff --git a/pnpm-lock.yaml b/pnpm-lock.yaml