Decode some html entities

src/constants/string_mappings.js

@@ -0,0 +1,37 @@
+/**
+* Copyright 2012-2016, Plotly, Inc.
+* All rights reserved.
+*
+* This source code is licensed under the MIT license found in the
+* LICENSE file in the root directory of this source tree.
+*/
+
+
+'use strict';
+
+// N.B. HTML entities are listed without the leading '&' and trailing ';'
+
+module.exports = {
+
+    entityToUnicode: {
+        'mu': 'μ',
+        'amp': '&',
+        'lt': '<',
+        'gt': '>',
+        'quot': '\"',
+        'nbsp': ' ',
+        'times': '×',
+        'plusmn': '±',
+        'deg': '°'
+    },
+
+    unicodeToEntity: {
+        '&': 'amp',
+        '<': 'lt',
+        '>': 'gt',
+        '"': 'quot',
+        '\'': '#x27',
+        '\/': '#x2F'
+    }
+
+};

src/lib/html2unicode.js

@@ -10,13 +10,7 @@
 'use strict';
 
 var toSuperScript = require('superscript-text');
-
-var ENTITIES = {
-    'mu': 'μ',
-    'amp': '&',
-    'lt': '<',
-    'gt': '>'
-};
+var stringMappings = require('../constants/string_mappings');
 
 function fixSuperScript(x) {
     var idx = 0;
@@ -40,6 +34,7 @@ function stripTags(x) {
 }
 
 function fixEntities(x) {
+    var entityToUnicode = stringMappings.entityToUnicode;
     var idx = 0;
 
     while((idx = x.indexOf('&', idx)) >= 0) {
@@ -49,7 +44,7 @@ function fixEntities(x) {
             continue;
         }
 
-        var entity = ENTITIES[x.slice(idx + 1, nidx)];
+        var entity = entityToUnicode[x.slice(idx + 1, nidx)];
         if(entity) {
             x = x.slice(0, idx) + entity + x.slice(nidx + 1);
         } else {

src/lib/svg_text_utils.js

@@ -15,6 +15,7 @@ var d3 = require('d3');
 
 var Lib = require('../lib');
 var xmlnsNamespaces = require('../constants/xmlns_namespaces');
+var stringMappings = require('../constants/string_mappings');
 
 // Append SVG
 
@@ -67,6 +68,7 @@ exports.convertToTspans = function(_context, _callback) {
     var str = _context.text();
     var converted = convertToSVG(str);
     var that = _context;
+
     // Until we get tex integrated more fully (so it can be used along with non-tex)
     // allow some elements to prohibit it by attaching 'data-notex' to the original
     var tex = (!that.attr('data-notex')) && converted.match(/([^$]*)([$]+[^$]*[$]+)([^$]*)/);
@@ -233,22 +235,48 @@ var PROTOCOLS = ['http:', 'https:', 'mailto:'];
 
 var STRIP_TAGS = new RegExp('</?(' + Object.keys(TAG_STYLES).join('|') + ')( [^>]*)?/?>', 'g');
 
+var ENTITY_TO_UNICODE = Object.keys(stringMappings.entityToUnicode).map(function(k) {
+    return {
+        regExp: new RegExp('&' + k + ';', 'g'),
+        sub: stringMappings.entityToUnicode[k]
+    };
+});
+
+var UNICODE_TO_ENTITY = Object.keys(stringMappings.unicodeToEntity).map(function(k) {
+    return {
+        regExp: new RegExp(k, 'g'),
+        sub: '&' + stringMappings.unicodeToEntity[k] + ';'
+    };
+});
+
 exports.plainText = function(_str) {
     // strip out our pseudo-html so we have a readable
     // version to put into text fields
     return (_str || '').replace(STRIP_TAGS, ' ');
 };
 
+function replaceFromMapObject(_str, list) {
+    var out = _str || '';
+
+    for(var i = 0; i < list.length; i++) {
+        var item = list[i];
+        out = out.replace(item.regExp, item.sub);
+    }
+
+    return out;
+}
+
+function convertEntities(_str) {
+    return replaceFromMapObject(_str, ENTITY_TO_UNICODE);
+}
+
 function encodeForHTML(_str) {
-    return (_str || '').replace(/&/g, '&amp;')
-        .replace(/</g, '&lt;')
-        .replace(/>/g, '&gt;')
-        .replace(/"/g, '&quot;')
-        .replace(/'/g, '&#x27;')
-        .replace(/\//g, '&#x2F;');
+    return replaceFromMapObject(_str, UNICODE_TO_ENTITY);
 }
 
 function convertToSVG(_str) {
+    _str = convertEntities(_str);
+
     var result = _str
         .split(/(<[^<>]*>)/).map(function(d) {
             var match = d.match(/<(\/?)([^ >]*)\s*(.*)>/i),
@@ -267,6 +295,7 @@ function convertToSVG(_str) {
                      * resurrect it.
                      */
                     extraStyle = extra.match(/^style\s*=\s*"([^"]+)"\s*/i);
+
                 // anchor and br are the only ones that don't turn into a tspan
                 if(tag === 'a') {
                     if(close) return '</a>';