From e058faf93dfc1ca6e1e693163efe52a6336b2b9a Mon Sep 17 00:00:00 2001
From: Espen Klem <espen.klem@gmail.com>
Date: Thu, 7 Mar 2019 15:15:27 +0100
Subject: [PATCH] Adding Swahili stopword list

---
 README.md           |  1 +
 lib/stopword.js     |  1 +
 lib/stopwords_he.js |  2 +-
 lib/stopwords_sw.js | 32 ++++++++++++++++++++++++++++++++
 package-lock.json   |  2 +-
 package.json        |  2 +-
 test/test.js        |  7 +++++++
 7 files changed, 44 insertions(+), 3 deletions(-)
 create mode 100644 lib/stopwords_sw.js

diff --git a/README.md b/README.md
index f3c1b2c..6e361b3 100644
--- a/README.md
+++ b/README.md
@@ -71,6 +71,7 @@ Arrays of stopwords for the following languages are supplied:
 * `pa_in` - Punjabi Gurmukhi
 * `ru` - Russian
 * `sv` - Swedish
+* `sw` - Swahili
 * `zh` - Chinese Simplified
 
 ```javascript
diff --git a/lib/stopword.js b/lib/stopword.js
index 25a072d..25e6f66 100644
--- a/lib/stopword.js
+++ b/lib/stopword.js
@@ -31,4 +31,5 @@ exports.pl = require('./stopwords_pl.js').words
 exports.pt = require('./stopwords_pt.js').words
 exports.ru = require('./stopwords_ru.js').words
 exports.sv = require('./stopwords_sv.js').words
+exports.sw = require('./stopwords_sw.js').words
 exports.zh = require('./stopwords_zh.js').words
diff --git a/lib/stopwords_he.js b/lib/stopwords_he.js
index aabc651..81c5892 100644
--- a/lib/stopwords_he.js
+++ b/lib/stopwords_he.js
@@ -199,7 +199,7 @@ var words = [
 "שם",
 "תהיה",
 "תחת"
-];
+]
 
 exports.words = words
 
diff --git a/lib/stopwords_sw.js b/lib/stopwords_sw.js
new file mode 100644
index 0000000..6f7e826
--- /dev/null
+++ b/lib/stopwords_sw.js
@@ -0,0 +1,32 @@
+/*
+The MIT License (MIT)
+
+
+Frequency sorted
+*/
+
+// Swahili - Stop words list. A list of commonly used words that have little
+// meaning and can be excluded from analysis.
+
+// This list is frequency sorted. That means it can be sliced from the bottom
+// and be less agressive in excluding stopwords */
+
+var words = ["na","ya","wa","kwa","ni","za","katika","la","kuwa","kama",
+  "kwamba","cha","hiyo","lakini","yake","hata","wakati","hivyo","sasa","wake",
+  "au","watu","hii","zaidi","vya","huo","tu","kwenye","si","pia","ili","moja",
+  "kila","baada","ambao","ambayo","yao","wao","kuna","hilo","kutoka","kubwa",
+  "pamoja","bila","huu","hayo","sana","ndani","mkuu","hizo","kufanya","wengi",
+  "hadi","mmoja","hili","juu","kwanza","wetu","kuhusu","baadhi","wote","yetu",
+  "hivi","kweli","mara","wengine","nini","ndiyo","zao","kati","hao","hapa",
+  "kutokana","muda","habari","ambaye","wenye","nyingine","hakuna","tena",
+  "hatua","bado","nafasi","basi","kabisa","hicho","nje","huyo","vile","yote",
+  "mkubwa","alikuwa","zote","leo","haya","huko","kutoa","mwa","kiasi","hasa",
+  "nyingi","kabla","wale","chini","gani","hapo","lazima","mwingine","bali",
+  "huku","zake","ilikuwa","tofauti","kupata","mbalimbali","pale","kusema",
+  "badala","wazi","yeye","alisema","hawa","ndio","hizi","tayari","wala",
+  "muhimu","ile","mpya","ambazo","dhidi","kwenda","sisi","kwani","jinsi",
+  "binafsi","kutumia","mbili","mbali","kuu","mengine","mbele","namna","mengi",
+  "upande"]
+
+exports.words = words
+
diff --git a/package-lock.json b/package-lock.json
index bb082c0..a826e04 100644
--- a/package-lock.json
+++ b/package-lock.json
@@ -1,6 +1,6 @@
 {
   "name": "stopword",
-  "version": "0.1.13",
+  "version": "0.1.14",
   "lockfileVersion": 1,
   "requires": true,
   "dependencies": {
diff --git a/package.json b/package.json
index 64279bd..f4a673e 100644
--- a/package.json
+++ b/package.json
@@ -1,6 +1,6 @@
 {
   "name": "stopword",
-  "version": "0.1.14",
+  "version": "0.1.15",
   "description": "A module for node.js that takes in text and returns text that is stripped of stopwords. Has pre-defined stopword lists for 20 languages and also takes lists with custom stopwords as input.",
   "main": "lib/stopword.js",
   "scripts": {
diff --git a/test/test.js b/test/test.js
index d85495d..a5531e2 100644
--- a/test/test.js
+++ b/test/test.js
@@ -145,6 +145,13 @@ describe('general stopwordiness:', function () {
     const newString = sw.removeStopwords(oldString, sw.my)
     newString.should.eql([''])
   })
+
+  it('should remove swahili stopwords', function () {
+    const oldString = 'kila mtu anaweza kuhariri makala yoyote kutoa makosa ya lugha kutohoa maneno na kuendeleza na kukuza makala kwa kuandika kwa ufupi au kwa urefu'.split(' ')
+    const newString = sw.removeStopwords(oldString, sw.sw)
+    newString.should.eql(['mtu','anaweza','kuhariri','makala','yoyote','makosa','lugha','kutohoa','maneno','kuendeleza','kukuza','makala','kuandika','ufupi','urefu'])
+  })
+
   
   
   // Right to Left languages