html-to-text - npm Package Compare versions

Comparing version 3.2.0 to 3.3.0

CHANGELOG.md

 # Changelog
 
+## Version 3.3.0
+
+* Ability to pass custom formatting via the `format` option #128
+* Enhanced support for alpha ordered list types added #123
+
 ## Version 3.2.0
@@ -4,0 +9,0 @@

lib/formatter.js

@@ -79,11 +79,13 @@ var _ = require('underscore');
     if (href) {
-      if (options.linkHrefBaseUrl && href.indexOf('/') === 0) {
-        href = options.linkHrefBaseUrl + href;
-      }
-      if (!options.hideLinkHrefIfSameAsText || href !== _s.replaceAll(result, '\n', '')) {
-        if (!options.noLinkBrackets) {
-          result += ' [' + href + ']';
-        } else {
-          result += ' ' + href;
+      if ((!options.noAnchorUrl) || (options.noAnchorUrl && href.indexOf('#') === -1)) {
+        if (options.linkHrefBaseUrl && href.indexOf('/') === 0) {
+          href = options.linkHrefBaseUrl + href;
         }
+        if (!options.hideLinkHrefIfSameAsText || href !== _s.replaceAll(result, '\n', '')) {
+          if (!options.noLinkBrackets) {
+            result += ' [' + href + ']';
+          } else {
+            result += ' ' + href;
+          }
+        }
       }
@@ -135,13 +137,18 @@ }
   // Return different functions for different OL types
-  var typeFunctions = {
-    1: function(start, i) { return i + 1 + start},
-    a: function(start, i) { return String.fromCharCode(i + start + 97)},
-    A: function(start, i) { return String.fromCharCode(i + start + 65)}
-  };
-  // Determine type
-  var olType = elem.attribs.type || '1'
+  var typeFunction = (function() {
+    // Determine type
+    var olType = elem.attribs.type || '1';
+    // TODO Imeplement the other valid types
+    //   Fallback to type '1' function for other valid types
+    switch(olType) {
+      case 'a': return function(start, i) { return String.fromCharCode(i + start + 97)};
+      case 'A': return function(start, i) { return String.fromCharCode(i + start + 65)};
+      case '1':
+      default: return function(start, i) { return i + 1 + start};
+    }
+  }())
   // Make sure there are list items present
   if (nonWhiteSpaceChildren.length) {
     // Calculate initial start from ol attribute
-    var start = Number(elem.attribs.start || '1') - 1
+    var start = Number(elem.attribs.start || '1') - 1;
     // Calculate the maximum length to i.
@@ -151,6 +158,6 @@ var maxLength = (nonWhiteSpaceChildren.length + start).toString().length;
       // Use different function depending on type
-      var index = typeFunctions[olType](start, i);
+      var index = typeFunction(start, i);
       // Calculate the needed spacing for nice indentation.
       var spacing = maxLength - index.toString().length;
-      var prefix = (olType === '1') ? ' ' + index + '. ' + _s.repeat(' ', spacing) : index + '. ';
+      var prefix = ' ' + index + '. ' + _s.repeat(' ', spacing);
       result += formatListItem(prefix, elem, fn, options);
@@ -157,0 +164,0 @@ });

lib/html-to-text.js

@@ -9,3 +9,3 @@ var fs = require('fs');
 var helper = require('./helper');
-var format = require('./formatter');
+var defaultFormat = require('./formatter');
 
@@ -29,4 +29,6 @@ // Which type of tags should not be parsed
     noLinkBrackets: false,
+    noAnchorUrl: true,
     baseElement: 'body',
     returnDomByDefault: true,
+    format: {},
     decodeOptions: {
@@ -112,2 +114,4 @@ isAttributeValue: false,
   var whiteSpaceRegex = /\s$/;
+  var format = _.assign({}, defaultFormat, options.format);
+
   _.each(dom, function(elem) {
@@ -114,0 +118,0 @@ switch(elem.type) {

package.json

 {
   "name": "html-to-text",
-  "version": "3.2.0",
+  "version": "3.3.0",
   "description": "Advanced html to plain text converter",
@@ -5,0 +5,0 @@ "main": "index.js",

README.md

@@ -74,3 +74,25 @@ # html-to-text
    * `forceWrapOnLimit` defines whether to break long words on the limit if `true`.
+ * `format` pass an object to enable custom formatting for specific elements (see below)
 
+### Override formatting for specific elements
+
+By using the `format` option, you can specify formatting for these elements: `text`, `image`, `lineBreak`, `paragraph`, `anchor`, `heading`, `table`, `orderedList`, `unorderedList`, `listItem`, `horizontalLine`.
+
+Each key must be a function which eventually receive `node` (the current node), `fn` (the next formatting function) and `options` (the options passed to html-to-text).
+
+```js
+var htmlToText = require('html-to-text');
+
+var text = htmlToText.fromString('<h1>Hello World</h1>', {
+	format: {
+		heading: function (node, fn, options) {
+			var h = fn(elem.children, options);
+			return '====\n' + h.toUpperCase() + '\n====';
+		}
+	}
+});
+
+console.log(text);
+```
+
 ## Command Line Interface
@@ -77,0 +99,0 @@

test/html-to-text.js

@@ -209,2 +209,16 @@ /* eslint max-len: "off" */
     });
+
+    it('should not return link for anchor if noAnchorUrl is set to true', function () {
+      var result = htmlToText.fromString('<a href="#link">test</a>', {
+        noAnchorUrl: true
+      });
+      expect(result).to.equal('test');
+    });
+
+    it('should return link for anchor if noAnchorUrl is set to false', function () {
+      var result = htmlToText.fromString('<a href="#link">test</a>', {
+        noAnchorUrl: false
+      });
+      expect(result).to.equal('test [#link]');
+    });
   });
@@ -241,5 +255,10 @@
 
+      it('should fallback to type="!" behavior if type attribute is invalid', function() {
+        var testString = '<ol type="1"><li>foo</li><li>bar</li></ol>';
+        expect(htmlToText.fromString(testString)).to.equal('1. foo\n 2. bar');
+      });
+
       it('should support the ordered list type="a" attribute', function() {
         var testString = '<ol type="a"><li>foo</li><li>bar</li></ol>';
-        expect(htmlToText.fromString(testString)).to.equal('a. foo\nb. bar');
+        expect(htmlToText.fromString(testString)).to.equal('a. foo\n b. bar');
       });
@@ -249,5 +268,19 @@
         var testString = '<ol type="A"><li>foo</li><li>bar</li></ol>';
-        expect(htmlToText.fromString(testString)).to.equal('A. foo\nB. bar');
+        expect(htmlToText.fromString(testString)).to.equal('A. foo\n B. bar');
       });
 
+      it('should support the ordered list type="i" attribute by falling back to type="1"', function() {
+        var testString = '<ol type="i"><li>foo</li><li>bar</li></ol>';
+        // TODO Implement lowercase roman numerals
+        // expect(htmlToText.fromString(testString)).to.equal('i. foo\nii. bar');
+        expect(htmlToText.fromString(testString)).to.equal('1. foo\n 2. bar');
+      });
+
+      it('should support the ordered list type="I" attribute by falling back to type="1"', function() {
+        var testString = '<ol type="I"><li>foo</li><li>bar</li></ol>';
+        // TODO Implement uppercase roman numerals
+        // expect(htmlToText.fromString(testString)).to.equal('I. foo\nII. bar');
+        expect(htmlToText.fromString(testString)).to.equal('1. foo\n 2. bar');
+      });
+
       it('should support the ordered list start attribute', function() {
@@ -322,2 +355,16 @@ var testString = '<ol start="2"><li>foo</li><li>bar</li></ol>';
 
+  describe('custom formatting', function () {
+    it('should allow to pass custom formatting functions', function () {
+      var result = htmlToText.fromString('<h1>TeSt</h1>', {
+        format: {
+          heading: function (elem, fn, options) {
+            var h = fn(elem.children, options);
+            return '====\n' + h.toLowerCase() + '\n====';
+          }
+        }
+      });
+      expect(result).to.equal('====\ntest\n====');
+    })
+  });
+
   describe('Base element', function () {
@@ -324,0 +371,0 @@ it('should retrieve and convert the entire document under `body` by default', function(done) {

New alerts

License Policy Violation

License

This package is not allowed per your license policy. Review the package's license to ensure compliance.

Found 1 instance in 1 package

Fixed alerts

License Policy Violation

License

This package is not allowed per your license policy. Review the package's license to ensure compliance.

Found 1 instance in 1 package

Improved metrics

Total package byte prevSize

84329

increased by4.12%

Lines of code

1078

increased by4.97%

Number of lines in readme file

332

increased by7.1%

No dependency changes