prop.put("viewMode_html", 1);
prop.put("viewMode_html_url", url.toNormalform(false, true));
}
} else if (viewMode.equals("parsed") || viewMode.equals("sentences") || viewMode.equals("words") || viewMode.equals("links")) {
// parsing the resource content
Document document = null;
try {
document = Document.mergeDocuments(response.url(), response.getMimeType(), response.parse());
if (document == null) {
prop.put("error", "5");
prop.put("error_errorText", "Unknown error");
prop.put("viewMode", VIEW_MODE_NO_TEXT);
return prop;
}
} catch (final Parser.Failure e) {
prop.put("error", "5");
prop.putHTML("error_errorText", e.getMessage());
prop.put("viewMode", VIEW_MODE_NO_TEXT);
return prop;
}
if (viewMode.equals("parsed")) {
final String content = UTF8.String(document.getTextBytes());
// content = wikiCode.replaceHTML(content); // added by Marc Nause
prop.put("viewMode", VIEW_MODE_AS_PARSED_TEXT);
prop.put("viewMode_title", document.dc_title());
prop.put("viewMode_creator", document.dc_creator());
prop.put("viewMode_subject", document.dc_subject(','));
prop.put("viewMode_description", document.dc_description());
prop.put("viewMode_publisher", document.dc_publisher());
prop.put("viewMode_format", document.dc_format());
prop.put("viewMode_identifier", document.dc_identifier());
prop.put("viewMode_source", url.toString());
prop.put("viewMode_lat", document.lat());
prop.put("viewMode_lon", document.lon());
prop.put("viewMode_parsedText", markup(wordArray, content).replaceAll("\n", "<br />").replaceAll("\t", " "));
} else if (viewMode.equals("sentences")) {
prop.put("viewMode", VIEW_MODE_AS_PARSED_SENTENCES);
final Collection<StringBuilder> sentences = document.getSentences(pre);
boolean dark = true;
int i = 0;
String sentence;
if (sentences != null) {
// Search word highlighting
for (final StringBuilder s: sentences) {
sentence = s.toString();
if (sentence.trim().length() > 0) {
prop.put("viewMode_sentences_" + i + "_nr", i + 1);
prop.put("viewMode_sentences_" + i + "_text", markup(wordArray, sentence));
prop.put("viewMode_sentences_" + i + "_dark", dark ? "1" : "0");
dark = !dark;
i++;
}
}
}
prop.put("viewMode_sentences", i);
} else if (viewMode.equals("words")) {
prop.put("viewMode", VIEW_MODE_AS_PARSED_WORDS);
final Collection<StringBuilder> sentences = document.getSentences(pre);
boolean dark = true;
int i = 0;
String sentence;
StringBuilder token;
if (sentences != null) {
// Search word highlighting
for (final StringBuilder s: sentences) {
sentence = s.toString();
Enumeration<StringBuilder> tokens = null;
tokens = new WordTokenizer(new ByteArrayInputStream(UTF8.getBytes(sentence)), LibraryProvider.dymLib);
while (tokens.hasMoreElements()) {
token = tokens.nextElement();
if (token.length() > 0) {
prop.put("viewMode_words_" + i + "_nr", i + 1);
prop.put("viewMode_words_" + i + "_word", token);
prop.put("viewMode_words_" + i + "_dark", dark ? "1" : "0");
dark = !dark;
i++;
}
}
}
}
prop.put("viewMode_words", i);
} else if (viewMode.equals("links")) {
prop.put("viewMode", VIEW_MODE_AS_LINKLIST);
boolean dark = true;
int i = 0;
i += putMediaInfo(prop, wordArray, i, document.getVideolinks(), "video", (i % 2 == 0), document.getAnchors());
i += putMediaInfo(prop, wordArray, i, document.getAudiolinks(), "audio", (i % 2 == 0), document.getAnchors());
dark = (i % 2 == 0);
final Map<MultiProtocolURI, ImageEntry> ts = document.getImages();
final Iterator<ImageEntry> tsi = ts.values().iterator();
ImageEntry entry;
while (tsi.hasNext()) {
entry = tsi.next();
prop.put("viewMode_links_" + i + "_nr", i);
prop.put("viewMode_links_" + i + "_dark", dark ? "1" : "0");
prop.put("viewMode_links_" + i + "_type", "image");
prop.put("viewMode_links_" + i + "_text", (entry.alt().isEmpty()) ? " " : markup(wordArray, entry.alt()));
prop.put("viewMode_links_" + i + "_url", entry.url().toNormalform(false, true));
prop.put("viewMode_links_" + i + "_link", markup(wordArray, entry.url().toNormalform(false, true)));
if (entry.width() > 0 && entry.height() > 0) {
prop.put("viewMode_links_" + i + "_rel", entry.width() + "x" + entry.height() + " Pixel");
} else {
prop.put("viewMode_links_" + i + "_rel", "");
}
prop.put("viewMode_links_" + i + "_name", "");
dark = !dark;
i++;
}
i += putMediaInfo(prop, wordArray, i, document.getApplinks(), "app", (i % 2 == 0), document.getAnchors());
i += putMediaInfo(prop, wordArray, i, document.getHyperlinks(), "link", (i % 2 == 0), document.getAnchors());
prop.put("viewMode_links", i);
}
if (document != null) document.close();
}
prop.put("error", "0");
prop.put("error_url", url.toNormalform(false, true));
prop.put("error_hash", urlHash);
prop.put("error_wordCount", wordCount);