sbd - npm Package Compare versions

Comparing version 0.0.2 to 0.0.3

data/abbr.js

@@ -0,0 +0,0 @@ exports.abbreviations = [

lib/Match.js

@@ -24,2 +24,3 @@
 // Starting a new sentence if beginning with capital letter
+// Exception: The word is enclosed in brackets
 exports.is_concatenated = function(word) {
@@ -31,3 +32,4 @@ var i = 0;
 
-        if (c === c.toUpperCase()) {
+        // Check if the next word starts with a letter
+        if (c.match(/[a-zA-Z].*/)) {
             return [word.slice(0, i), word.slice(i+1)];
@@ -34,0 +36,0 @@ }

lib/tokenizer.js

@@ -36,3 +36,6 @@ /*jshint node:true, laxcomma:true */
 
-        if (Match.is_boundary_char(words[i]) || String.ends_with_char(words[i], "?!") || words[i] === newline_placeholder_t) {
+        if (Match.is_boundary_char(words[i])      ||
+            String.ends_with_char(words[i], "?!") ||
+            words[i] === newline_placeholder_t)
+        {
             if (newline_boundary) {
@@ -117,3 +120,5 @@ current.pop();
         // Single words, could be "enumeration lists"
-        if (sentences[i].length === 1 && sentences[i][0].length < 4 && sentences[i][0].indexOf('.') > -1) {
+        if (sentences[i].length === 1 && sentences[i][0].length < 4 &&
+            sentences[i][0].indexOf('.') > -1)
+        {
             // Check if there is a next sentence
@@ -131,2 +136,2 @@ // It should not be another list item
     return result;
-};
+};

package.json

 {
   "name": "sbd",
-  "version": "0.0.2",
+  "version": "0.0.3",
   "description": "Split text into sentences",
@@ -5,0 +5,0 @@ "main": "lib/tokenizer.js",

README.md

@@ -0,0 +0,0 @@ # Sentence Boundary Detection (SBD)

test/test.js

@@ -54,2 +54,11 @@ /*jshint node:true, laxcomma:true */
 
+    describe('Questionmark is skipped inside brackets', function () {
+        var entry = "A sentence [example?] that should not (Though sometimes...) be two or more (but one!) sentences.";
+        var sentences = tokenizer.sentences(entry);
+
+        it("should get 1 sentence", function () {
+            assert.equal(sentences.length, 1);
+        });
+    });
+
     describe('Skip abbreviations', function () {
@@ -56,0 +65,0 @@ var entry = "In I.C.T we have multiple challenges! There should only be two sentences.";

No alert changes

Improved metrics

Total package byte prevSize

26722

increased by8.27%

Lines of code

1447

increased by1.05%

No dependency changes