oxjs/source/Ox/js/HTML.js

'use strict';

(function() {

    var defaultTags = [
            // inline formatting
            'b', 'code', 'i', 's', 'sub', 'sup', 'u',
            // block formatting
            'blockquote', 'h1', 'h2', 'h3', 'p', 'pre',
            // lists
            'li', 'ol', 'ul',
            // tables
            'table', 'tbody', 'td', 'tfoot', 'th', 'thead', 'tr',
            // other
            'a', 'br', 'img',
            // special
            'rtl', '[]'
        ],
        htmlEntities = {
            '"': '&quot;', '&': '&amp;', "'": '&apos;', '<': '&lt;', '>': '&gt;'
        },
        regexp = {
            entity: /&[^\s]*;/g,
            tag: new RegExp('<\\/?(' + [
                'a', 'b', 'br', 'code', 'i', 's', 'span', 'u'
            ].join('|') + ')\\/?>', 'gi')
        },
        replace = {
            mail: [
                /\b([0-9A-Z\.\+\-_]+@(?:[0-9A-Z\-]+\.)+[A-Z]{2,6})\b/gi,
                '<a href="mailto:$1">$1</a>'
            ],
            namedEntity: [
                new RegExp('(' + Ox.values(htmlEntities).join('|') + ')', 'g'),
                function(match) {
                    return Ox.keyOf(htmlEntities, match);
                }
            ],
            numericEntity: [
                /&#([0-9A-FX]+);/gi,
                function(match, code) {
                    return Ox.char(
                        /^X/i.test(code)
                            ? parseInt(code.slice(1), 16)
                            : parseInt(code, 10)
                    );
                }
            ],
            tag: {
                a: [
                    [
                        /<a [^<>]*?href="((\/|https?:\/\/|mailto:).+?)".*?>/gi,
                        '<a href="{1}">',
                    ],
                    [
                        /<\/a>/gi,
                        '</a>'
                    ]
                ],
                img: [
                    [
                        /<img [^<>]*?src="((\/|https?:\/\/).+?)".*?>/gi,
                        '<img src="$1">'
                    ]
                ],
                rtl: [
                    [
                        /<rtl>/gi,
                        '<div style="direction: rtl">'
                    ],
                    [
                        /<\/rtl>/gi,
                        '</div>'
                    ]
                ],
                '*': function(tag) {
                    return [
                        [
                            new RegExp('</?' + tag + ' ?/?>', 'gi'),
                            '{0}'
                        ]
                    ];
                }
            },
            url: [
                /\b((https?:\/\/|www\.).+?)([\.,:;!\?\)\]]*?(\s|$))/gi,
                function(string, url, prefix, end) {
                    prefix = prefix.toLowerCase() == 'www.' ? 'http://' : '';
                    return Ox.formatString(
                        '<a href="{prefix}{url}">{url}</a>{end}',
                        {end: end, prefix: prefix, url: url}
                    );
                }
            ]
        },
        salt = Ox.range(2).map(function(){
            return Ox.range(16).map(function() {
                return Ox.char(65 + Ox.random(26));
            }).join('');
        });

    // Splits a string into text (even indices) and tags (odd indices), ignoring
    // tags with starting positions that are included in the ignore array
    function splitHTMLTags(string, ignore) {
        var isTag = false, ret = [''];
        ignore = ignore || [];
        Ox.forEach(string, function(char, i) {
            if (!isTag && char == '<' && ignore.indexOf(i) == -1) {
                isTag = true;
                ret.push('');
            }
            ret[ret.length - 1] += char;
            if (isTag && char == '>') {
                isTag = false;
                ret.push('');
            }
        });
        return ret;
    };

    /*@
    Ox.addLinks <f> Takes a string and adds links for e-mail addresses and URLs
        (string[, isHTML]) -> <s> Formatted string
        string <s> String
        isHTML <b|false> If true, ignore matches in tags or enclosed by links
        > Ox.addLinks('foo bar <foo@bar.com>')
        'foo bar &lt;<a href="mailto:foo@bar.com">foo@bar.com</a>&gt;'
        > Ox.addLinks('www.foo.com/bar#baz, etc.')
        '<a href="http://www.foo.com/bar#baz">www.foo.com/bar#baz</a>, etc.'
        > Ox.addLinks('<a href="http://foo.com">foo.com</a>', true)
        '<a href="http://foo.com">foo.com</a>'
    @*/
    Ox.addLinks = function(string, isHTML) {
        var isLink = false;
        function replaceString(string) {
            return string
                .replace(replace.mail[0], replace.mail[1])
                .replace(replace.url[0], replace.url[1]);
        }
        return isHTML
            ? splitHTMLTags(string).map(function(string, i) {
                var isTag = i % 2;
                if (isTag) {
                    if (/^<a/.test(string)) {
                        isLink = true;
                    } else if (/^<\/a/.test(string)) {
                        isLink = false;
                    }
                }
                return isTag || isLink ? string : replaceString(string); 
            }).join('')
            : Ox.normalizeHTML(replaceString(string));
    };

    /*@
    Ox.encodeHTMLEntities <f> Encodes HTML entities
        (string[, encodeAll]) -> <s> String
        string <s> String
        encodeAll <b|false> If true, encode characters > 127 as numeric entities
        > Ox.encodeHTMLEntities('<\'&"> äbçdê')
        '&lt;&apos;&amp;&quot;&gt; äbçdê'
        > Ox.encodeHTMLEntities('<\'&"> äbçdê', true)
        '&lt;&apos;&amp;&quot;&gt; &#x00E4;b&#x00E7;d&#x00EA;'
    @*/
    Ox.encodeHTMLEntities = function(string, encodeAll) {
        return Ox.map(String(string), function(char) {
            var code = char.charCodeAt(0);
            if (code < 128) {
                char = char in htmlEntities ? htmlEntities[char] : char;
            } else if (encodeAll) {
                char = '&#x' + Ox.pad(code.toString(16).toUpperCase(), 4) + ';';
            }
            return char;
        });
    };

    /*@
    Ox.decodeHTMLEntities <f> Decodes HTML entities
        (string[, decodeAll]) -> <s> String
        string <s> String
        decodeAll <b|false> If true, decode named entities for characters > 127
            Note that <code>decodeAll</code> relies on
            <code>Ox.normalizeHTML</code>, which uses the DOM and may transform
            the string
        > Ox.decodeHTMLEntities('&#x003C;&#x0027;&#x0026;&#x0022;&#x003E;')
        '<\'&">'
        > Ox.decodeHTMLEntities('&lt;&apos;&amp;&quot;&gt;')
        '<\'&">'
        > Ox.decodeHTMLEntities('&#x00E4;b&#x00E7;d&#x00EA;')
        'äbçdê'
        > Ox.decodeHTMLEntities('&auml;b&ccedil;d&ecirc;')
        '&auml;b&ccedil;d&ecirc;'
        > Ox.decodeHTMLEntities('&auml;b&ccedil;d&ecirc;', true)
        'äbçdê'
        > Ox.decodeHTMLEntities('<b>&beta;')
        '<b>&beta;'
        > Ox.decodeHTMLEntities('<b>&beta;', true)
        '<b>β</b>'
        > Ox.decodeHTMLEntities('&lt;b&gt;')
        '<b>'
    @*/
    Ox.decodeHTMLEntities = function(string, decodeAll) {
        return decodeAll
            ? Ox.decodeHTMLEntities(Ox.normalizeHTML(string))
            : String(string)
                .replace(replace.namedEntity[0], replace.namedEntity[1])
                .replace(replace.numericEntity[0], replace.numericEntity[1]);
    };

    /*@
    Ox.highlightHTML <f> Highlight matches in string
        (string, query, classname[, isHTML]) -> Output string
        string <s> Input string
        query <r|s> Case-insentitive query string, or regular expression
        classname <s> Class name for matches
        isHTML <b|false> If true, the input string is treated as HTML
        > Ox.highlightHTML('<name>', 'name', 'c')
        '&lt;<span class="c">name</span>&gt;'
        > Ox.highlightHTML('<span class="name">name</span>', 'name', 'c', true)
        '<span class="name"><span class="c">name</span></span>'
        > Ox.highlightHTML('amp &amp; amp', 'amp', 'c', true)
        '<span class="c">amp</span> &amp; <span class="c">amp</span>'
        > Ox.highlightHTML('amp &amp; amp', 'amp & amp', 'c', true)
        '<span class="c">amp &amp; amp</span>'
        > Ox.highlightHTML('<b>&lt;b&gt;</b>', '<b>', 'c', true)
        '<span class="c"><b>&lt;b&gt;</b></span>'
        > Ox.highlightHTML('<b>&lt;b&gt;</b>', '&lt;b&gt;', 'c', true)
        '<b>&lt;b&gt;</b>'
        > Ox.highlightHTML('foo<b>bar</b>baz', 'foobar', 'c', true)
        '<span class="c">foo<b>bar</b></span>baz'
        > Ox.highlightHTML('foo<p>bar</p>baz', 'foobar', 'c', true)
        'foo<p>bar</p>baz'
        > Ox.highlightHTML('foo <br/>bar baz', 'foo bar', 'c', true)
        '<span class="c">foo <br>bar</span> baz'
    @*/
    Ox.highlightHTML = function(string, query, classname, isHTML) {
        var cursor = 0,
            entities = [],
            matches = [],
            re = Ox.isRegExp(query) ? query
                : new RegExp(Ox.escapeRegExp(query), 'gi'),
            span = ['<span class="' + classname + '">', '</span>'],
            tags = [];
        function insert(array) {
            // for each replacement
            array.forEach(function(v) {
                // replace the modified value with the original value
                string = Ox.splice(string, v.position, v.length, v.value);
                // for each match
                matches.forEach(function(match) {
                    if (v.position < match.position) {
                        // replacement is before match, update match position
                        match.position += v.value.length - v.length;
                    } else if (
                        v.position < match.position + match.value.length
                    ) {
                        // replacement is inside match, update match value
                        match.value = Ox.splice(
                            match.value, v.position - match.position, v.length,
                            v.value
                        );
                    }
                });
            });
        }
        if (isHTML) {
            string = Ox.normalizeHTML(string)
                // remove inline tags
                .replace(regexp.tag, function(value, tag, position) {
                    tags.push({
                        length: 0, position: position, value: value
                    });
                    return '';
                })
                // decode html entities
                .replace(regexp.entity, function(value, position) {
                    var ret = Ox.decodeHTMLEntities(value, true);
                    entities.push({
                        length: ret.length, position: position, value: value
                    });
                    return ret;
                });
            // if decoding entities has created new tags, ignore them
            splitHTMLTags(string, entities.map(function(entity) {
                return entity.position;
            })).forEach(function(v, i) {
                if (i % 2 == 0) {
                    // outside tags, find matches and save position and value
                    v.replace(re, function(value, position) {
                        matches.push(
                            {position: cursor + position, value: value}
                        );
                    });
                }
                cursor += v.length;
            });
            insert(entities);
            insert(tags);
            // for each match (in reverse order, so that positions are correct)
            matches.reverse().forEach(function(match) {
                // wrap it in a span
                string = Ox.splice(
                    string, match.position, match.value.length,
                    span.join(match.value)
                );
            });
            
        } else {
            string = string.replace(re, function(value) {
                return span.join(value);
            });
        }
        // if isHTML, we may have enclosed single opening or closing tags in a 
        // span, if not isHTML, the string may contain '<', '>' or '&', so in
        // both cases, we have to normalize
        return Ox.normalizeHTML(string);
    };

    /*@
    Ox.normalizeHTML <f> Normalize HTML (using the DOM)
        > Ox.normalizeHTML('<b>foo')
        '<b>foo</b>'
        > Ox.normalizeHTML('<b>foo</b></b>')
        '<b>foo</b>'
        > Ox.normalizeHTML('&lt;&apos;&amp;&quot;&gt; &#x00E4;b&#x00E7;d&#x00EA;')
        '&lt;\'&amp;"&gt; äbçdê'
    @*/
    Ox.normalizeHTML = function(html) {
        return Ox.$('<div>').html(html).html();
    };

    /*@
    Ox.sanitizeHTML <f> Takes untrusted HTML and returns something trustworthy
        > Ox.sanitizeHTML('http://foo.com, bar')
        '<a href="http://foo.com">http://foo.com</a>, bar'
        > Ox.sanitizeHTML('http://foo.com/foo?bar, bar')
        '<a href="http://foo.com/foo?bar">http://foo.com/foo?bar</a>, bar'
        > Ox.sanitizeHTML('(see: www.foo.com)')
        '(see: <a href="http://www.foo.com">www.foo.com</a>)'
        > Ox.sanitizeHTML('foo@bar.com')
        '<a href="mailto:foo@bar.com">foo@bar.com</a>'
        > Ox.sanitizeHTML('<a href="mailto:foo@bar.com">foo</a>')
        '<a href="mailto:foo@bar.com">foo</a>'
        > Ox.sanitizeHTML('<a href="http://foo.com">foo</a>')
        '<a href="http://foo.com">foo</a>'
        > Ox.sanitizeHTML('<a href="http://foo.com" onclick="alert()">foo</a>')
        '<a href="http://foo.com">foo</a>'
        > Ox.sanitizeHTML('<a href="javascript:alert()">foo</a>')
        '&lt;a href="javascript:alert()"&gt;foo'
        > Ox.sanitizeHTML('<a href="foo">foo</a>')
        '&lt;a href="foo"&gt;foo'
        > Ox.sanitizeHTML('<a href="/foo">foo</a>')
        '<a href="/foo">foo</a>'
        > Ox.sanitizeHTML('[http://foo.com foo]')
        '<a href="http://foo.com">foo</a>'
        > Ox.sanitizeHTML('<rtl>foo</rtl>')
        '<div style="direction: rtl">foo</div>'
        > Ox.sanitizeHTML('<script>alert()</script>')
        '&lt;script&gt;alert()&lt;/script&gt;'
        > Ox.sanitizeHTML('\'foo\' < \'bar\' && "foo" > "bar"')
        '\'foo\' &lt; \'bar\' &amp;&amp; "foo" &gt; "bar"'
        > Ox.sanitizeHTML('<b>foo')
        '<b>foo</b>'
        > Ox.sanitizeHTML('<b>foo</b></b>')
        '<b>foo</b>'
    @*/
    Ox.sanitizeHTML = function(html, tags) {
        var matches = [];
        tags = tags || defaultTags;
        // html = Ox.clean(html); fixme: can this be a parameter?
        if (tags.indexOf('[]') > -1) {
            html = html.replace(/\[((\/|https?:\/\/|mailto:).+?) (.+?)\]/gi, '<a href="$1">$3</a>');
            tags = tags.filter(function(tag) {
                return tag != '[]';
            });
        }
        tags.forEach(function(tag) {
            var array = replace.tag[tag] || replace.tag['*'](tag);
            Ox.forEach(array, function(value) {
                html = html.replace(value[0], function() {
                    matches.push(Ox.formatString(value[1], arguments));
                    return salt.join(matches.length - 1);
                });
            });
        });
        html = Ox.addLinks(Ox.encodeHTMLEntities(html), true);
        matches.forEach(function(match, i) {
            html = html.replace(new RegExp(salt.join(i)), match);
        });
        html = html.replace(/\n\n/g, '<br/><br/>');
        // Close extra opening and remove extra closing tags.
        // Note: this converts '&apos;' to "'" and '&quot;' to '"'
        return Ox.normalizeHTML(html);
    };

}());