degausser - npm Package Compare versions

Comparing version 2.3.2 to 2.4.0

dist/degausser.js

@@ -294,3 +294,38 @@ (function (global, factory) {
   };
+  /**
+   * Gets the alt text from an element, if it exists, otherwise returns placeholder alt text composed of 100 unit separator character.
+   * If node has empty alt attribute or alt attribute with empty string, this will return the placeholder alt text instead.
+   * @param node
+   * @param placeholderCharacter
+   * @param placeholderLength
+   * @returns {string}
+   */
 
+
+  var getAltText = function getAltText(node, placeholderCharacter, placeholderLength) {
+    var altText = node.getAttribute('alt');
+
+    if (!altText) {
+      var altTextPlaceholder = placeholderCharacter.repeat(placeholderLength);
+      return altTextPlaceholder;
+    }
+
+    return altText;
+  };
+  /**
+   * Checks if element with given tagname can have an alt attribute.
+   * @param tagName
+   * @returns {boolean}
+   */
+
+
+  var elementCanHaveAltText = function elementCanHaveAltText(tagName) {
+    if (!tagName) {
+      return false;
+    }
+
+    var tagNameLowerCase = tagName.toLowerCase();
+    return tagNameLowerCase === 'img' || tagNameLowerCase === 'image' || tagNameLowerCase === 'area';
+  };
+
   var StringCollector = /*#__PURE__*/function () {
@@ -399,5 +434,6 @@ function StringCollector() {
 
-        if (node.hasAttribute('alt')) {
+        if (elementCanHaveAltText(node.tagName)) {
           this.processBreaks();
-          this.text.push(" ".concat(node.getAttribute('alt'), " "));
+          var altText = getAltText(node, this.options.placeholderString, this.options.placeholderCopies);
+          this.text.push(" ".concat(altText, " "));
           return true;
@@ -649,7 +685,8 @@ }
 
-        if (node.hasAttribute('alt')) {
+        if (elementCanHaveAltText(node.tagName)) {
           this.processBreaks();
+          var altText = getAltText(node, this.options.placeholderString, this.options.placeholderCopies);
           this.text.push({
             node: node,
-            string: " ".concat(node.getAttribute('alt'), " ")
+            string: " ".concat(altText, " ")
           });
@@ -741,4 +778,11 @@ return true;
                 if (entity.node.nodeType === Node.TEXT_NODE || entity.node.tagName === 'img') {
-                  var nodeContent = entity.node.tagName === 'img' ? entity.node.getAttribute('alt').normalize() : entity.node.textContent.normalize();
+                  var nodeContent = void 0;
 
+                  if (elementCanHaveAltText(entity.node.tagName)) {
+                    var altText = getAltText(entity.node, this.options.placeholderString, this.options.placeholderCopies).normalize();
+                    nodeContent = altText;
+                  } else {
+                    nodeContent = entity.node.textContent.normalize();
+                  }
+
                   for (var charInMap = 0, charInNode = 0; charInNode < nodeContent.length; ++charInNode) {
@@ -851,8 +895,24 @@ var isEqual = entity.content.charAt(charInMap) === nodeContent.charAt(charInNode);
 
+  /**
+   * Extracts text from the given node.
+   * Options include (but are not limited to):
+   * - placeholderString: string to take the place of alt text when alt it is empty/undefined
+   * - placeholderCopies: the number of times placeholderString repeats
+   * @param parentNode
+   * @param options
+   * @returns {*}
+   */
+
   var degausser = function degausser(parentNode) {
     var options = arguments.length > 1 && arguments[1] !== undefined ? arguments[1] : {};
-    var collector = new StringCollector(options);
+    var unitSeparatorCode = 31;
+    var defaultOptions = {
+      placeholderString: String.fromCharCode(unitSeparatorCode),
+      placeholderCopies: 100
+    };
+    var finalOptions = Object.assign(defaultOptions, options);
+    var collector = new StringCollector(finalOptions);
 
-    if (options.map) {
-      collector = new MapCollector(options);
+    if (finalOptions.map) {
+      collector = new MapCollector(finalOptions);
     }
@@ -865,2 +925,3 @@
     var map = arguments.length > 3 && arguments[3] !== undefined ? arguments[3] : null;
+    var options = arguments.length > 4 && arguments[4] !== undefined ? arguments[4] : {};
     var docType = doc.nodeType;
@@ -873,5 +934,5 @@
     if (map === null) {
-      map = degausser(doc, {
-        map: true
-      });
+      var finalOptions = Object.assign({}, options);
+      finalOptions.map = true;
+      map = degausser(doc, finalOptions);
     }
@@ -878,0 +939,0 @@

package.json

 {
     "name": "degausser",
-    "version": "2.3.2",
+    "version": "2.4.0",
     "description": "Transforms HTML to plain text by eliminating tags from a document.",
@@ -5,0 +5,0 @@ "author": "FlowPub",

src/degausser.js

@@ -5,7 +5,23 @@ import { StringCollector } from './stringCollector'
 
+/**
+ * Extracts text from the given node.
+ * Options include (but are not limited to):
+ * - placeholderString: string to take the place of alt text when alt it is empty/undefined
+ * - placeholderCopies: the number of times placeholderString repeats
+ * @param parentNode
+ * @param options
+ * @returns {*}
+ */
 export const degausser = (parentNode, options = {}) => {
-  let collector = new StringCollector(options)
+  const unitSeparatorCode = 31
+  const defaultOptions = {
+    placeholderString: String.fromCharCode(unitSeparatorCode),
+    placeholderCopies: 100,
+  }
+  const finalOptions = Object.assign(defaultOptions, options)
 
-  if (options.map) {
-    collector = new MapCollector(options)
+  let collector = new StringCollector(finalOptions)
+
+  if (finalOptions.map) {
+    collector = new MapCollector(finalOptions)
   }
@@ -16,3 +32,3 @@
 
-export const getRangeFromOffset = (start, end, doc = document, map = null) => {
+export const getRangeFromOffset = (start, end, doc = document, map = null, options = {}) => {
   const docType = doc.nodeType
@@ -27,3 +43,5 @@ if (
   if (map === null) {
-    map = degausser(doc, { map: true })
+    const finalOptions = Object.assign({}, options)
+    finalOptions.map = true
+    map = degausser(doc, finalOptions)
   }
@@ -30,0 +48,0 @@

src/mapCollector.js

@@ -9,2 +9,4 @@ import {
   isElementBlacklisted,
+  getAltText,
+  elementCanHaveAltText,
 } from './util'
@@ -185,6 +187,8 @@
 
-    if (node.hasAttribute('alt')) {
+    if (elementCanHaveAltText(node.tagName)) {
       this.processBreaks()
-      this.text.push({ node, string: ` ${node.getAttribute('alt')} ` })
 
+      const altText = getAltText(node, this.options.placeholderString, this.options.placeholderCopies)
+      this.text.push({ node, string: ` ${altText} ` })
+
       return true
@@ -269,6 +273,13 @@ }
           ) {
-            const nodeContent =
-              entity.node.tagName === 'img'
-                ? entity.node.getAttribute('alt').normalize()
-                : entity.node.textContent.normalize()
+            let nodeContent
+            if (elementCanHaveAltText(entity.node.tagName)) {
+              const altText = getAltText(
+                entity.node,
+                this.options.placeholderString,
+                this.options.placeholderCopies
+              ).normalize()
+              nodeContent = altText
+            } else {
+              nodeContent = entity.node.textContent.normalize()
+            }
 
@@ -275,0 +286,0 @@ for (

src/stringCollector.js

@@ -8,2 +8,4 @@ import {
   isElementBlacklisted,
+  getAltText,
+  elementCanHaveAltText,
 } from './util'
@@ -116,6 +118,12 @@
 
-    if (node.hasAttribute('alt')) {
+    if (elementCanHaveAltText(node.tagName)) {
       this.processBreaks()
-      this.text.push(` ${node.getAttribute('alt')} `)
 
+      const altText = getAltText(
+        node,
+        this.options.placeholderString,
+        this.options.placeholderCopies
+      )
+      this.text.push(` ${altText} `)
+
       return true
@@ -122,0 +130,0 @@ }

src/util.js

@@ -217,2 +217,34 @@ function autoBind() {
 
+/**
+ * Gets the alt text from an element, if it exists, otherwise returns placeholder alt text composed of 100 unit separator character.
+ * If node has empty alt attribute or alt attribute with empty string, this will return the placeholder alt text instead.
+ * @param node
+ * @param placeholderCharacter
+ * @param placeholderLength
+ * @returns {string}
+ */
+const getAltText = (node, placeholderCharacter, placeholderLength) => {
+  const altText = node.getAttribute('alt')
+  if (!altText) {
+    const altTextPlaceholder = placeholderCharacter.repeat(placeholderLength)
+    return altTextPlaceholder
+  }
+
+  return altText
+}
+
+/**
+ * Checks if element with given tagname can have an alt attribute.
+ * @param tagName
+ * @returns {boolean}
+ */
+const elementCanHaveAltText = (tagName) => {
+  if (!tagName) {
+    return false
+  }
+
+  const tagNameLowerCase = tagName.toLowerCase()
+  return tagNameLowerCase === 'img' || tagNameLowerCase === 'image' || tagNameLowerCase === 'area'
+}
+
 export {
@@ -227,2 +259,4 @@ autoBind,
   isCharWhitespace,
+  getAltText,
+  elementCanHaveAltText,
 }

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

53616

increased by9.02%

Lines of code

1627

increased by7.68%

No dependency changes