public static Status getStatusTweet(String JSONString) {
if (JSONString == null) return null;
Status tweet = null;
try {
// parse json to object type
tweet = DataObjectFactory.createStatus(JSONString);
} catch (TwitterException e) {
System.err.println("error parsing tweet object");
return null;
}
return tweet;
}
public void testUnparsable() throws Exception {
String str;
str = "";
try {
DataObjectFactory.createStatus(str);
fail("should fail");
} catch (TwitterException expected) {
} catch (Error notExpected) {
fail("failed" + notExpected.getMessage());
}
try {
DataObjectFactory.createStatus(str);
fail("should fail");
} catch (TwitterException expected) {
} catch (Error notExpected) {
fail("failed" + notExpected.getMessage());
}
str =
"{\"in_reply_to_status_id_str\":null,\"place\":null,\"in_reply_to_user_id\":null,\"text\":\"working\",\"contributors\":null,\"retweet_count\":0,\"in_reply_to_user_id_str\":null,\"retweeted\":false,\"id_str\":\"794626207\",\"source\":\"\\u003Ca href=\\\"http:\\/\\/twitterhelp.blogspot.com\\/2008\\/05\\/twitter-via-mobile-web-mtwittercom.html\\\" rel=\\\"nofollow\\\"\\u003Emobile web\\u003C\\/a\\u003E\",\"truncated\":false,\"geo\":null,\"in_reply_to_status_id\":null,\"favorited\":false,\"user\":{\"show_all_inline_media\":false,\"geo_enabled\":false,\"profile_background_tile\":false,\"time_zone\":null,\"favourites_count\":0,\"description\":null,\"friends_count\":0,\"profile_link_color\":\"0084B4\",\"location\":null,\"profile_sidebar_border_color\":\"C0DEED\",\"id_str\":\"14481043\",\"url\":null,\"follow_request_sent\":false,\"statuses_count\":1,\"profile_use_background_image\":true,\"lang\":\"en\",\"profile_background_color\":\"C0DEED\",\"profile_image_url\":\"http:\\/\\/a3.twimg.com\\/a\\/1292975674\\/images\\/default_profile_3_normal.png\",\"profile_background_image_url\":\"http:\\/\\/a3.twimg.com\\/a\\/1292975674\\/images\\/themes\\/theme1\\/bg.png\",\"followers_count\":44,\"protected\":false,\"contributors_enabled\":false,\"notifications\":false,\"screen_name\":\"Yusuke\",\"name\":\"Yusuke\",\"is_translator\":false,\"listed_count\":1,\"following\":false,\"verified\":false,\"profile_text_color\":\"333333\",\"id\":14481043,\"utc_offset\":null,\"created_at\":\"Tue Apr 22 21:49:13 +0000 2008\",\"profile_sidebar_fill_color\":\"DDEEF6\"},\"id\":794626207,\"coordinates\":null,\"in_reply_to_screen_name\":null,\"created_at\":\"Tue Apr 2200 21:49:34 +0000 2008\"";
try {
DataObjectFactory.createCategory(str);
fail("should fail");
} catch (TwitterException expected) {
expected.printStackTrace();
} catch (Error notExpected) {
fail("failed" + notExpected.getMessage());
}
try {
DataObjectFactory.createCategory(str);
fail("should fail");
} catch (TwitterException expected) {
} catch (Error notExpected) {
fail("failed" + notExpected.getMessage());
}
}
public static void main(String argv[]) throws IOException {
argv[0] = "-mis";
argv[1] = "/Users/Indri/Eclipse_workspace/GeoNames/cities1000.txt";
argv[2] = "/Users/Indri/Eclipse_workspace/GazIndex";
argv[3] = "SampleInput/jsonTweets.txt";
argv[4] = "-json";
argv[5] = "SampleOutput/jsonTweets.out.csv";
boolean misspell = argv[0].equals("-mis") ? true : false;
String dicPath = argv[1]; // = "GeoNames/allCountries.txt";// gazetteer from geonames
String indexPath = argv[2]; // index path
String input = argv[3]; // = "tweet.csv";//to be determined.// test file path
String type = argv[4]; // -json or -text
String output = argv[5]; // = "output2.csv"; //output file path
CollaborativeIndex ci =
new CollaborativeIndex()
.config("GazIndex/StringIndex", "GazIndex/InfoIndex", "mmap", "mmap")
.open();
EnglishParser enparser = new EnglishParser("res/", ci, false);
ContextDisamb c = new ContextDisamb();
LangDetector lang = new LangDetector("res/langdetect.profile");
BufferedReader reader = GetReader.getUTF8FileReader(argv[3]);
CsvWriter writer = new CsvWriter(output, ',', Charset.forName("utf-8")); // write
writer.writeRecord(new String[] {"SPANISH TWEETS", "LOCATIONS"});
String line = null;
while ((line = reader.readLine()) != null) {
line = line.trim();
if (line.length() == 0) continue;
Tweet t = new Tweet();
String text = null;
if (argv[4].equals("-text")) text = line;
else
try {
text = (DataObjectFactory.createStatus(line.trim()).getText());
} catch (TwitterException e) {
// TODO Auto-generated catch block
System.err.println("JSON format corrupted, or no content.");
continue;
}
t.setText(text);
List<String> match = enparser.parse(t);
// Generate Matches
if (match == null || match.size() == 0) {
/** write blank result and the line itself if no match found. */
writer.writeRecord(new String[] {text, ""});
continue;
}
HashSet<String> reducedmatch = new HashSet<String>();
for (String s : match) reducedmatch.add(s.substring(3, s.length() - 3));
// Disambiguate topo
HashMap<String, String[]> result = c.returnBestTopo(ci, reducedmatch);
if (result == null) {
System.out.println("No GPS for any location is found.");
} else {
System.out.println("The grounded location(s) are:");
String topoStr = "";
for (String topo : result.keySet())
topoStr +=
"["
+ (topo
+ ": "
+ result.get(topo)[2]
+ " "
+ result.get(topo)[0]
+ " "
+ result.get(topo)[1])
+ "] ";
writer.writeRecord(new String[] {text, topoStr});
}
}
reader.close();
writer.close();
}
public JSONTweet readLine() {
String line;
try {
line = br.readLine();
} catch (IOException e1) {
// TODO Auto-generated catch block
e1.printStackTrace();
return null;
}
if (line == null) return null;
Status tweet = null;
try {
// parse json to object type
tweet = DataObjectFactory.createStatus(line);
} catch (TwitterException e) {
System.err.println("error parsing tweet object");
return null;
}
jsontweet.JSON = line;
jsontweet.id = tweet.getId();
jsontweet.source = tweet.getSource();
jsontweet.text = tweet.getText();
jsontweet.createdat = tweet.getCreatedAt();
jsontweet.tweetgeolocation = tweet.getGeoLocation();
User user;
if ((user = tweet.getUser()) != null) {
jsontweet.userdescription = user.getDescription();
jsontweet.userid = user.getId();
jsontweet.userlanguage = user.getLang();
jsontweet.userlocation = user.getLocation();
jsontweet.username = user.getName();
jsontweet.usertimezone = user.getTimeZone();
jsontweet.usergeoenabled = user.isGeoEnabled();
if (user.getURL() != null) {
String url = user.getURL().toString();
jsontweet.userurl = url;
String addr = url.substring(7).split("/")[0];
String[] countrysuffix = addr.split("[.]");
String suffix = countrysuffix[countrysuffix.length - 1];
jsontweet.userurlsuffix = suffix;
try {
InetAddress address = null; // InetAddress.getByName(user.getURL().getHost());
String generate_URL
// =
// "http://www.geobytes.com/IpLocator.htm?GetLocation&template=php3.txt&IpAddress="
= "http://www.geoplugin.net/php.gp?ip=" + address.getHostAddress();
URL data = new URL(generate_URL);
URLConnection yc = data.openConnection();
BufferedReader in = new BufferedReader(new InputStreamReader(yc.getInputStream()));
String inputLine;
String temp = "";
while ((inputLine = in.readLine()) != null) {
temp += inputLine + "\n";
}
temp = temp.split("s:2:\"")[1].split("\"")[0];
jsontweet.userurllocation = temp;
} catch (Exception uhe) {
// uhe.printStackTrace();
jsontweet.userurllocation = null;
}
}
}
return jsontweet;
}
