NEW: allow characters adjacent to brackets within words (#18)

closes #18
digitallinguistics · Jul 7, 2020 · 9e47c1d · 9e47c1d
1 parent 9ebcd2f
commit 9e47c1d
Show file tree

Hide file tree

Showing 2 changed files with 11 additions and 9 deletions.
diff --git a/alignWords.js b/alignWords.js
@@ -12,7 +12,7 @@ function alignWords(lines, options = {}) {
 
   const [leftBracket, rightBracket] = groupWith.trim().split(``);
 
-  let wordRegExp = /(?<bracketed>\[.*?\])|(?<unbracketed>[^\s]+)/gu;
+  let wordRegExp = /(?<bracketed>\S*\[.*?\]\S*)|(?<unbracketed>\S+)/gu;
 
   if (groupWith !== `[]`) {
 

diff --git a/alignWords.test.js b/alignWords.test.js
@@ -33,7 +33,7 @@ describe(`alignWords`, function() {
 
   });
 
-  it.only(`aligns by character length`, function() {
+  it(`aligns by character length`, function() {
 
     const lines = [
       `cafȩ́ word t̓aatn̓a ʔunaakʔaɬ`,
@@ -55,22 +55,24 @@ describe(`alignWords`, function() {
 
   });
 
-  it(`groups words with [brackets]`, function() {
+  it.only(`groups words with [brackets]`, function() {
 
     const lines = [
-      `waxdungu qasi [qapx cuyi]`,
-      `waxt-qungu qasi [qapx cuy-i]`,
-      `day-one man he.came`,
+      `word word [word word] [stem]-suffix prefix-[word word] [word word],`,
+      `m-m m [m m-m] [m]-m m-[m m] [m m]`,
+      `gl-gl gl gl gl-gl gl-[gl gl] gl`,
     ];
 
     const expectedResult = [
-      `waxdungu   qasi [qapx cuyi]`,
-      `waxt-qungu qasi [qapx cuy-i]`,
-      `day-one    man  he.came`,
+      `word  word [word word] [stem]-suffix prefix-[word word] [word word],`,
+      `m-m   m    [m m-m]     [m]-m         m-[m m]            [m m]`,
+      `gl-gl gl   gl          gl-gl         gl-[gl gl]         gl`,
     ];
 
     const aligned = alignWords(lines);
 
+    console.log(JSON.stringify(aligned, null, 2));
+
     aligned.forEach((line, i) => {
 
       expect(line).to.be(expectedResult[i]);