Split hyphens in a match the same way the tokenizer does

src/1-one/match/methods/parseMatch/03-splitHyphens.js

@@ -1,16 +1,12 @@
-const hasDash = /[a-z0-9][-–—][a-z]/i
+import { hasHyphen } from '../../../tokenize/methods/02-terms/01-hyphens.js'
 
 // match 're-do' -> ['re','do']
+// split them the same way the tokenizer does - 'x-ray' stays one word
 const splitHyphens = function (regs, world) {
-  const prefixes = world.model.one.prefixes
   for (let i = regs.length - 1; i >= 0; i -= 1) {
     const reg = regs[i]
-    if (reg.word && hasDash.test(reg.word)) {
+    if (reg.word && hasHyphen(reg.word, world.model)) {
       let words = reg.word.split(/[-–—]/g)
-      // don't split 're-cycle', etc
-      if (Object.hasOwn(prefixes, words[0])) {
-        continue
-      }
       words = words.filter(w => w).reverse()
       regs.splice(i, 1)
       words.forEach(w => {

tests/one/tokenize/hyphen-matrix.test.js

@@ -79,7 +79,13 @@ test('match-dash', function (t) {
     'fifty-six bottles',
     'a 10-minute speech',
     'self-serve',
-    'non-sequitur'
+    'non-sequitur',
+    'an x-ray was done',
+    'the x-ray-machine broke',
+    'she had covid-19',
+    'a t-shirt',
+    'a flower-like shape',
+    'the president-elect',
   ]
   arr.forEach(str => {
     t.equal(nlp(str).has(str), true, here + '[dash] ' + str)