com.offerready.xslt.HtmlBodyExtractor Maven / Gradle / Ivy

Go to download
Show more of this group Show more artifacts with this name
Show all versions of xslt-library Show documentation
Utility classes for XSLT processing used at Offer-Ready GmbH
The newest version!
package com.offerready.xslt;

import com.databasesandlife.util.Timer;

import javax.annotation.CheckForNull;
import javax.annotation.Nonnull;
import java.util.ArrayList;
import java.util.List;
import java.util.regex.Pattern;

/**
 * Extracts the body of an HTML report document.
 *    
 * An "HTML document generator" generates HTML, suitable for viewing in a browser, or inclusion in an iframe. 
 * However, this HTML contains e.g. <html> tags which are not suitable for display within a <div>.
 * Objects of this class extract the relevant information from the HTML document, and return an HTML
 * string suitable for inclusion in a <div>.
 *    
 * The entire contents of the <body> are returned.
 * Any <style> tags in the <head> are extracted are prepended to this body.
 */
public class HtmlBodyExtractor {
    
    protected List scriptsToIgnore = new ArrayList<>();
    
    protected static class Range {
        int startIncl, endExcl;
        Range(int s, int e) { startIncl = s; endExcl = e; }
        boolean overlaps(Range x) { 
            if (x.endExcl <= startIncl) return false;
            if (x.startIncl >= endExcl) return false;
            return true;
        }
    }
    
    /**
     * For example if JQuery should not be included; simply add "jquery" to this method and this <script>
     * tag will not be included in the result.
     */
    public @Nonnull HtmlBodyExtractor addScriptToIgnore(@Nonnull String scriptSubstring) {
        scriptsToIgnore.add(scriptSubstring);
        return this;
    }
    
    protected @Nonnull String ignoreScripts(@Nonnull String html) {
        for (var s : scriptsToIgnore)
            html = html.replaceAll(
                "", 
                "");
        return html;
    }
    
    protected void extractElements(
        @Nonnull StringBuilder result, @Nonnull String input,
        @Nonnull List ranges,
        @Nonnull String start, @CheckForNull String endOfStartOrNull, @Nonnull String end
    ) {
        var startIdx = -1;
        var ourResult = new StringBuilder();
        while ((startIdx = input.indexOf(start, startIdx+1)) >= 0) {
            var idxOfEndTag = input.indexOf(end, startIdx);
            Range range = endOfStartOrNull == null 
                ? new Range(startIdx, idxOfEndTag + end.length())
                : new Range(input.indexOf(endOfStartOrNull, startIdx) + endOfStartOrNull.length(), idxOfEndTag);
            var rangeOverlaps = false;
            for (var r : ranges) if (r.overlaps(range)) rangeOverlaps = true;
            if (rangeOverlaps) continue;
            ourResult.append(ignoreScripts(input.substring(range.startIncl, range.endExcl)));
            ranges.add(range);
        }
        ourResult.append("\n");
        result.insert(0, ourResult);
    }
    
    public @Nonnull String extractBody(@Nonnull String htmlText) {
        try (var ignored = new Timer("HtmlBodyExtractor.extractBody")) {
            var result  = new StringBuilder();  // 2.5 seconds at 1k iterations on i3
            var ranges = new ArrayList();
            
            extractElements(result, htmlText, ranges, "",  "");
            extractElements(result, htmlText, ranges, "");
            extractElements(result, htmlText, ranges, "");
            extractElements(result, htmlText, ranges, "");
                
            return result.toString();
        }
    }
}