public String processVoiceResults(ArrayList<String> matchedStrings) {
Log.d(TAG, "Initial results = " + matchedStrings);
// Default result is the most likely match of those returned
String result = matchedStrings.get(0);
boolean resultFound = false;
for (int i = 0; i < Search.getCurrentSearch().getSearchObjects().size() && !resultFound; i++) {
LocalEntity currentSearchItem = Search.getCurrentSearch().getSearchObjects().get(i);
for (int j = 0; j < matchedStrings.size() && !resultFound; j++) {
for (int k = 0; k < currentSearchItem.getSearchTerms().size() && !resultFound; k++) {
// Log.d(TAG, "matched string = " + matchedStrings.get(j));
// Log.d(TAG, "search string = " +
// currentSearchItem.getSearchTerms().get(k));
// If magic algorithm finds a match
if (StringUtils.getLevenshteinDistance(
matchedStrings.get(j), currentSearchItem.getSearchTerms().get(k))
< (currentSearchItem.getSearchTerms().get(k).length() / 3)
|| currentSearchItem.getSearchTerms().get(k).equals(matchedStrings.get(j))) {
resultFound = true;
if (Search.getCurrentSearch().getName().equals("yesno")) {
Log.d(TAG, "got here");
result = currentSearchItem.getSearchTerms().get(k);
} else {
result = Search.getCurrentSearch().getSearchObjects().get(i).getName();
}
Log.d(TAG, "result = " + result);
// If result found then update next search based on match
processExpectedResult(result);
return null;
}
}
}
}
return processUnexpectedResult(result);
}
/**
* Returns the nearest valid dictionary words to a given word, by calculating the Levenshtein
* distance between the given word and all valid ones, and returning an array of tiered sizes,
* based on the length of the given word. The returned array is ordered by closeness of match, and
* contains more close matches than distant ones. It also checks for more matches with larger
* Levenshtein distances for long given words than for short ones.
*
* @param wordToMatch the word to find similar valid words for.
* @return an array of valid dictionary words close to the given one, ordered by relevance and in
* order of increasing Levenshtein distance.
*/
public static String[] closeMatches(String wordToMatch) {
String[] dictionary = VALUES.toArray(new String[VALUES.size()]);
int count = 0;
int countOne = 0;
int countTwo = 0;
int countThree = 0;
String[] nearestMatches = new String[15];
String[] nearestMatchesTwo = new String[10];
String[] nearestMatchesThree = new String[5];
for (int i = 0; i < dictionary.length; i++) {
if (count < 30
&& StringUtils.getLevenshteinDistance(wordToMatch, dictionary[i]) == 1
&& countOne < 15) {
nearestMatches[countOne] = dictionary[i];
countOne++;
count++;
} else if (count < 30
&& StringUtils.getLevenshteinDistance(wordToMatch, dictionary[i]) == 2
&& countTwo < 10) {
nearestMatchesTwo[countTwo] = dictionary[i];
countTwo++;
count++;
} else if (count < 30
&& StringUtils.getLevenshteinDistance(wordToMatch, dictionary[i]) == 3
&& countThree < 5) {
nearestMatchesThree[countThree] = dictionary[i];
countThree++;
count++;
}
}
String[] moreMatches = ArrayUtils.addAll(nearestMatches, nearestMatchesTwo);
String[] evenMoreMatches = ArrayUtils.addAll(moreMatches, nearestMatchesThree);
if (wordToMatch.length() > 5) {
return removeNullVals(moreMatches);
}
if (wordToMatch.length() > 9) {
return removeNullVals(evenMoreMatches);
}
return removeNullVals(nearestMatches);
}
/**
* Checks the given player name.
*
* @param expected expected name
* @param actual actual name
*/
private void assertPlayerName(String expected, String actual) {
// OCR'ing player names is hard, so we're going to give Tesseract a break and allow for some
// mismatched
// characters.
int distance = StringUtils.getLevenshteinDistance(expected, actual);
String message =
String.format(
"Expected \"%s\", got \"%s\", Levenshtein distance: %s",
expected, actual, "" + distance);
log.info(message);
Assert.assertTrue(message, distance < 5);
}
public static SupportedWebDriver getClosestDriverValueTo(final String value) {
SupportedWebDriver closestDriver = null;
int closestDriverDistance = Integer.MAX_VALUE;
for (SupportedWebDriver supportedDriver : values()) {
int distance = StringUtils.getLevenshteinDistance(supportedDriver.toString(), value);
if (distance < closestDriverDistance) {
closestDriverDistance = distance;
closestDriver = supportedDriver;
}
}
return closestDriver;
}
private static boolean containsAll(List<String> what, List<String> where) {
int count = 0;
for (String whatC : what) {
for (String whereC : where) {
if (StringUtils.getLevenshteinDistance(whatC, whereC) <= Math.max(1, whereC.length() / 3)
&& Math.abs(whatC.length() - whereC.length()) <= 1) {
count++;
}
}
}
return (what.size() <= 3) ? what.size() <= count : what.size() - 1 <= count;
}
public static List<Weibo> filterWeibo(List<Weibo> topicList) throws IOException {
for (int i = 0; i < topicList.size(); i++) {
String str1 = topicList.get(i).getContent();
for (int j = i + 1; j < topicList.size(); j++) {
String str2 = topicList.get(j).getContent();
int ld = StringUtils.getLevenshteinDistance(str1, str2);
double sim = 1 - (ld / (double) Math.max(str1.length(), str2.length()));
if ((sim > 0.4) || (str2.length() < 5)) {
topicList.remove(j);
j--;
}
}
}
return topicList;
}
/**
* Finds and returns the string from the alternatives that closest matches the input string based
* on the Levenshtein distance.
*
* @see <a href="http://en.wikipedia.org/wiki/Levenshtein_distance">Levenshtein Distance</a>
* @param input
* @param alternatives
* @return the string from the alternatives that closest matches the input string based on the
* Levenshtein distance.
*/
public static String getClosestMatchingString(String input, Collection<String> alternatives) {
if (input == null || alternatives == null || alternatives.isEmpty()) {
return null;
}
String current = null;
int minDistance = Integer.MAX_VALUE;
for (String alternative : alternatives) {
int currentDistance = StringUtils.getLevenshteinDistance(input, alternative);
if ((current == null) || (currentDistance < minDistance)) {
current = alternative;
minDistance = currentDistance;
}
}
return current;
}
private static Player getPlayerByDisplayNameAlt(String needle) {
needle = needle.toLowerCase().trim();
Integer minEditDistance = null;
Player minEditMatch = null;
for (Player player : Bukkit.getOnlinePlayers()) {
String haystack = player.getDisplayName().toLowerCase().trim();
int editDistance = StringUtils.getLevenshteinDistance(needle, haystack.toLowerCase());
if (minEditDistance == null || minEditDistance.intValue() > editDistance) {
minEditDistance = editDistance;
minEditMatch = player;
}
}
return minEditMatch;
}
/**
* Gets map object associated with given map name.
*
* @param name name of map
* @return map object associated with the name
*/
public static AutoRefMap getMap(String name) {
// assume worldName exists
if (name == null) return null;
name = AutoRefMatch.normalizeMapName(name);
// if there is no map library, quit
File mapLibrary = AutoRefMap.getMapLibrary();
if (!mapLibrary.exists()) return null;
AutoRefMap bmap = null;
int ldist = MAX_NAME_DISTANCE;
for (AutoRefMap map : getAvailableMaps()) {
String mapName = AutoRefMatch.normalizeMapName(map.name);
int namedist = StringUtils.getLevenshteinDistance(name, mapName);
if (namedist <= ldist) {
bmap = map;
ldist = namedist;
}
}
// get best match
return bmap;
}
public static Movie scrapeMovie(
File movieFile,
SiteParsingProfile siteToParseFrom,
String urlToScrapeFromDMM,
boolean useURLtoScrapeFrom)
throws IOException {
// System.out.println("movieFile = " + movieFile);
String searchString = siteToParseFrom.createSearchString(movieFile);
SearchResult[] searchResults = null;
int searchResultNumberToUse = 0;
// no URL was passed in so we gotta figure it ourselves
if (!useURLtoScrapeFrom) {
searchResults = siteToParseFrom.getSearchResults(searchString);
int levDistanceOfCurrentMatch = 999999; // just some super high number
String idFromMovieFile = SiteParsingProfile.findIDTagFromFile(movieFile);
// loop through search results and see if URL happens to contain ID number in the URL. This
// will improve accuracy!
for (int i = 0; i < searchResults.length; i++) {
String urltoMatch = searchResults[i].getUrlPath().toLowerCase();
String idFromMovieFileToMatch = idFromMovieFile.toLowerCase().replaceAll("-", "");
// System.out.println("Comparing " + searchResults[i].toLowerCase() + " to " +
// idFromMovieFile.toLowerCase().replaceAll("-", ""));
if (urltoMatch.contains(idFromMovieFileToMatch)) {
// let's do some fuzzy logic searching to try to get the "best" match in case we got some
// that are pretty close
// and update the variables accordingly so we know what our best match so far is
int candidateLevDistanceOfCurrentMatch =
StringUtils.getLevenshteinDistance(urltoMatch.toLowerCase(), idFromMovieFileToMatch);
if (candidateLevDistanceOfCurrentMatch < levDistanceOfCurrentMatch) {
levDistanceOfCurrentMatch = candidateLevDistanceOfCurrentMatch;
searchResultNumberToUse = i;
}
}
}
}
// just use the URL to parse from the parameter
else if (useURLtoScrapeFrom) {
searchResults = new SearchResult[1];
if (siteToParseFrom instanceof DmmParsingProfile)
searchResults[0] = new SearchResult(urlToScrapeFromDMM);
else if (siteToParseFrom instanceof Data18MovieParsingProfile
|| siteToParseFrom instanceof Data18WebContentParsingProfile)
searchResults[0] = new SearchResult(urlToScrapeFromDMM);
else if (siteToParseFrom instanceof JavLibraryParsingProfile)
searchResults[0] =
new SearchResult(
((JavLibraryParsingProfile) siteToParseFrom).getOverrideURLJavLibrary());
else if (siteToParseFrom instanceof IAFDParsingProfile)
searchResults[0] = new SearchResult(urlToScrapeFromDMM);
}
if (searchResults != null
&& searchResults.length > 0
&& searchResults[searchResultNumberToUse].getUrlPath().length() > 0) {
System.out.println(
"Scraping this webpage for movie: "
+ searchResults[searchResultNumberToUse].getUrlPath());
// for now just set the movie to the first thing found unless we found a link which had
// something close to the ID
Document searchMatch =
Jsoup.connect(searchResults[searchResultNumberToUse].getUrlPath())
.timeout(0)
.userAgent("Mozilla/5.0 (Windows NT 6.1; WOW64; rv:5.0) Gecko/20100101 Firefox/5.0")
.get();
siteToParseFrom.setDocument(searchMatch);
siteToParseFrom.setOverrideURLDMM(urlToScrapeFromDMM);
return new Movie(siteToParseFrom);
} else return null; // TODO return some kind of default movie
}
@Override
public double compare(String word1, String word2) {
double distance = StringUtils.getLevenshteinDistance(word1, word2);
distance /= word1.length() > word2.length() ? word1.length() : word2.length();
return 1 - distance;
}
