/**
* Searches a target URL recursively for the target string, up to a max number of links and
* returns all the links containing the target string as a word in the link or in its content
* body.
*
* @param url The link to search from
* @param target The word to search for
* @param max Max number of links to open
* @return An array of URLs containing all the links and pages matching the search
*/
public URL[] search(URL url, String target, int max) {
this.max = max;
final ArrayList<URL> targetFoundAt = new ArrayList<>();
queue.enqueue(url);
while (!queue.isEmpty()) {
final URL link = queue.dequeue();
StdOut.println(discovered.size());
discovered.add(url);
if (this.scan(link, target)) {
targetFoundAt.add(link);
}
}
URL[] returnArray = new URL[targetFoundAt.size()];
return targetFoundAt.toArray(returnArray);
}
/**
* Scans a given file for an exact match to the word target. We have chosen to use equals and
* contains so that sites including the target in the URL get included.
*
* @param url The link to scan
* @param target The word to search for
* @return Whether or not this file contains the target word.
*/
public boolean scan(final URL url, final String target) {
boolean foundTarget = false;
In in = new In(url);
String[] strings;
try {
strings = in.readAllStrings();
} catch (NullPointerException e) {
return false;
}
for (String word : strings) {
// If the word can be identified as a link
if (word.startsWith(LINK_IDENTIFIER + "http")) {
/*
Prepares the link for construction into a URL. First it replaces the link identifier, typically
a string like href=", however this is only valid for the beginning. We later use a split to remove the
" following the actual URL, as well as possibly a tag closure like >. The reason for implementing this
with split, is that we want to preserve the word immediately following the tag closure if applicable.
This is because this may be a word matching target, so we later replace word with the result of that
split where applicable.
*/
String[] splits = word.replace(LINK_IDENTIFIER, "").split("\"(>|)");
if (splits[0] != null) {
splits[0] = splits[0].replace("&", "&");
try {
URL newLink = new URL(splits[0]);
// Don't enqueue the link if we've already discovered it.
if (!discovered.contains(newLink) && discovered.size() < max) {
queue.enqueue(newLink);
discovered.add(newLink);
}
} catch (MalformedURLException ignored) {
// We have already turned the string in question into a valid link. It should not be
// identified
// as a link if it does not conform to the above statements. It could still be saved
// with more
// complex parsing, but we don't wish to expend resources parsing javascripts or by
// analysing
// patterns to determine what to fix.
}
}
if (splits.length > 1 && splits[1] != null) {
word = splits[1];
}
}
// Make sure we check for lower case values when matching words.
if (word.toLowerCase().equals(target) || word.toLowerCase().contains(target)) {
foundTarget = true;
}
}
return foundTarget;
}