used docpa to parse and traverse

package.json

@@ -1,6 +1,6 @@
 {
   "name": "jsonfromtable",
-  "version": "3.0.0",
+  "version": "3.1.0",
   "description": "Convert html tables to object (or array)",
   "main": "dist/index.js",
   "scripts": {
@@ -37,9 +37,7 @@
   },
   "homepage": "https://github.com/coderosh/jsonfromtable#readme",
   "dependencies": {
-    "css-select": "^4.2.0",
-    "dom-serializer": "^1.3.2",
-    "htmlparser2": "^7.2.0",
+    "docpa": "^1.0.0",
     "node-fetch": "^2.6.1"
   },
   "gitHooks": {

src/utils.ts

@@ -1,27 +1,27 @@
-import render from 'dom-serializer'
-import { selectAll, selectOne } from 'css-select'
-import { DomUtils, parseDocument } from 'htmlparser2'
-
-const { getAttributeValue, textContent } = DomUtils
+import Docpa from 'docpa'
 
 function getRowWithColumns(
-  tableDoc: ReturnType<typeof parseDocument>,
+  tableDoc: Docpa,
   selectors: [string, string],
   shouldBeText: boolean,
   trim: boolean
 ) {
   const rowSelector = selectors[0] || 'tr'
   const colSelector = selectors[1] || 'td,th'
 
-  return selectAll(rowSelector, tableDoc).map((tr) =>
-    selectAll(colSelector, tr).map((td) => {
-      const rowspan = (td && +getAttributeValue(td as any, 'rowspan')!) || 1
-      const colspan = (td && +getAttributeValue(td as any, 'colspan')!) || 1
-      const value =
-        (td && (shouldBeText ? textContent(td) : render(td.children))) || ''
+  return tableDoc.querySelectorAll(rowSelector).map((tr) => {
+    if (!tr) return []
+    return tr.querySelectorAll(colSelector).map((td) => {
+      if (!td) return { value: '', colspan: 1, rowspan: 1 }
+
+      const rowspan = +(td.getAttribute('rowspan') || 1) || 1
+      const colspan = +(td.getAttribute('colspan') || 1) || 1
+
+      const value = (shouldBeText ? td.textContent : td.innerHTML) || ''
+
       return { value: trim ? value.trim() : value, colspan, rowspan }
     })
-  )
+  })
 }
 
 interface ParseTableOptions {
@@ -39,8 +39,8 @@ function parseTable(html: string, options: ParseTableOptions) {
     trim = true,
   } = options
 
-  const document = parseDocument(html)
-  const table = selectOne(tableSelector, document)
+  const document = new Docpa(html)
+  const table = document.querySelector(tableSelector)
 
   if (!table) throw new Error(`${tableSelector} not found in document.`)
 

yarn.lock

@@ -1247,10 +1247,10 @@ cross-spawn@^7.0.2, cross-spawn@^7.0.3:
     shebang-command "^2.0.0"
     which "^2.0.1"
 
-css-select@^4.2.0:
-  version "4.2.0"
-  resolved "https://registry.yarnpkg.com/css-select/-/css-select-4.2.0.tgz#ab28276d3afb00cc05e818bd33eb030f14f57895"
-  integrity sha512-6YVG6hsH9yIb/si3Th/is8Pex7qnVHO6t7q7U6TIUnkQASGbS8tnUDBftnPynLNnuUl/r2+PTd0ekiiq7R0zJw==
+css-select@^4.2.1:
+  version "4.2.1"
+  resolved "https://registry.yarnpkg.com/css-select/-/css-select-4.2.1.tgz#9e665d6ae4c7f9d65dbe69d0316e3221fb274cdd"
+  integrity sha512-/aUslKhzkTNCQUB2qTX84lVmfia9NyjP3WpDGtj/WxhwBzWBYUV3DgUpurHTme8UTPcPlAD1DJ+b0nN/t50zDQ==
   dependencies:
     boolbase "^1.0.0"
     css-what "^5.1.0"
@@ -1364,6 +1364,15 @@ dir-glob@^3.0.1:
   dependencies:
     path-type "^4.0.0"
 
+docpa@^1.0.0:
+  version "1.0.0"
+  resolved "https://registry.yarnpkg.com/docpa/-/docpa-1.0.0.tgz#98a27c64f30777ac169e494f271b12bf55b3bbde"
+  integrity sha512-v1aSAg6uLEAES4IIuFqQxl7+DrCpWkBf/8IJJo0vSU7FLMU6lTSzvizPvx0TSFgmsu93klxGDGlRIBKkJ4JDxA==
+  dependencies:
+    css-select "^4.2.1"
+    dom-serializer "^1.3.2"
+    htmlparser2 "^7.2.0"
+
 doctrine@^2.1.0:
   version "2.1.0"
   resolved "https://registry.yarnpkg.com/doctrine/-/doctrine-2.1.0.tgz#5cd01fc101621b42c4cd7f5d1a66243716d3f39d"