private String blankSectionHeaders(String markup, StringBuffer context) {
Pattern p = Pattern.compile("(={2,})([^=]+)\\1");
Matcher m = p.matcher(markup);
int lastPos = 0;
StringBuilder sb = new StringBuilder();
while (m.find()) {
sb.append(markup.substring(lastPos, m.start()));
sb.append(getSpaceString(m.group().length()));
String title = m.group(2).trim();
if (!title.equalsIgnoreCase("see also")
&& !title.equalsIgnoreCase("external links")
&& !title.equalsIgnoreCase("references")
&& !title.equalsIgnoreCase("further reading")) context.append("\n").append(title);
lastPos = m.end();
}
sb.append(markup.substring(lastPos));
return sb.toString();
}
private String blankLinks(String markup, StringBuffer context, HashSet<Integer> bannedTopics) {
List<Integer> linkStack = new ArrayList<Integer>();
Pattern p = Pattern.compile("(\\[\\[|\\]\\])");
Matcher m = p.matcher(markup);
StringBuilder sb = new StringBuilder();
int lastIndex = 0;
while (m.find()) {
String tag = markup.substring(m.start(), m.end());
if (tag.equals("[[")) linkStack.add(m.start());
else {
if (!linkStack.isEmpty()) {
int linkStart = linkStack.size() - 1;
linkStack.remove(linkStack.size() - 1);
if (linkStack.isEmpty()) {
sb.append(markup.substring(lastIndex, linkStart));
// we have the whole link, possibly with other links nested inside.
for (int i = linkStart; i < m.end(); i++) sb.append(" ");
processLink(markup.substring(linkStart + 2, m.start()), context, bannedTopics);
lastIndex = m.end();
}
}
}
}
if (!linkStack.isEmpty()) {
System.err.println(
"WikiPreprocessor| Warning: links were not well formed, so we cannot guarantee that they were stripped out correctly. ");
}
sb.append(markup.substring(lastIndex));
return sb.toString();
}
private String blankTemplates(String markup) {
List<Integer> templateStack = new ArrayList<Integer>();
Pattern p = Pattern.compile("(\\{\\{|\\}\\})");
Matcher m = p.matcher(markup);
StringBuilder sb = new StringBuilder();
int lastIndex = 0;
while (m.find()) {
String tag = markup.substring(m.start(), m.end());
if (tag.equals("{{")) templateStack.add(m.start());
else {
if (!templateStack.isEmpty()) {
int templateStart = templateStack.size() - 1;
templateStack.remove(templateStack.size() - 1);
if (templateStack.isEmpty()) {
sb.append(markup.substring(lastIndex, templateStart));
// we have the whole template, with other templates nested inside
for (int i = templateStart; i < m.end(); i++) sb.append(" ");
lastIndex = m.end();
}
}
}
}
if (!templateStack.isEmpty())
System.err.println(
"WikiPreprocessor | Warning: templates were not well formed, so we cannot guarantee that they were stripped out correctly. ");
sb.append(markup.substring(lastIndex));
return sb.toString();
}
private String clearAllMentionsRetainFirstCharacter(String regex, String text) {
Pattern p = Pattern.compile(regex, Pattern.CASE_INSENSITIVE + Pattern.DOTALL);
Matcher m = p.matcher(text);
int lastPos = 0;
StringBuilder sb = new StringBuilder();
while (m.find()) {
sb.append(text.substring(lastPos, m.start()));
sb.append(text.charAt(m.start()));
for (int i = 1; i < m.group().length(); i++) sb.append(" ");
lastPos = m.end();
}
sb.append(text.substring(lastPos));
return sb.toString();
}