diff options
author | S129778 <S129778@S129778.campus.tue.nl> | 2014-05-21 16:54:34 +0200 |
---|---|---|
committer | S129778 <S129778@S129778.campus.tue.nl> | 2014-05-21 16:54:34 +0200 |
commit | edbab8bc4e3236853fbf16ed1b65a262c178a1c2 (patch) | |
tree | 5d3803d216ded1a5f6ac63bb87a4d50f9433169c | |
parent | 4831e2b08b225a30d418197844f3d71e9e0c81cb (diff) | |
download | Goldfarmer-edbab8bc4e3236853fbf16ed1b65a262c178a1c2.tar.gz |
Change wordcloud with added brands
-rw-r--r-- | src/main/Analyzor.java | 581 |
1 files changed, 296 insertions, 285 deletions
diff --git a/src/main/Analyzor.java b/src/main/Analyzor.java index bbe1b5e..12a6a4c 100644 --- a/src/main/Analyzor.java +++ b/src/main/Analyzor.java @@ -1,285 +1,296 @@ -package main;
-
-import analysis.BrandChecker;
-import database.NamedPreparedStatement;
-import database.QueryUtils;
-import java.io.File;
-import java.io.FileNotFoundException;
-import java.io.IOException;
-import java.io.PrintWriter;
-import java.io.UnsupportedEncodingException;
-import java.sql.Connection;
-import java.sql.PreparedStatement;
-import java.sql.ResultSet;
-import java.sql.SQLException;
-import java.util.List;
-import java.util.HashMap;
-import java.util.Map.Entry;
-import java.util.Scanner;
-
-/**
- * The sentiment analysis class that rates tweets based on a unigram and bigram
- * set of weights.
- */
-public class Analyzor {
-
- /**
- * The map that matches single words to their weights.
- */
- private final HashMap<String, Double> unimap = new HashMap();
-
- /**
- * The map that matches word pairs to their weights.
- */
- private final HashMap<String, Double> bimap = new HashMap();
-
- /**
- * The results of a query, maybe return from query().
- */
- private ResultSet data;
-
- /**
- * The persistent connection to the database.
- */
- private final Connection connection;
-
- /**
- * @param connection An open connection to the database.
- */
- public Analyzor(Connection connection) {
- this.connection = connection;
- }
-
- /**
- * Read the unigram and bigram lexica.
- *
- * @throws FileNotFoundException
- */
- public void readLexicon() throws FileNotFoundException {
- if (!unimap.isEmpty()) {
- // data is already read.
- return;
- }
- System.err.println("Trying to read lexicons...");
- // A unigram is in the format (WS = whitespace):
- // word <WS> rating <WS> ??? <WS> ??
- // A bigram has an two WS-separated words instead of one.
- try (Scanner uniScanner = new Scanner(new File("unigrams-pmilexicon.txt"));
- Scanner biScanner = new Scanner(new File("bigrams-pmilexicon.txt"));) {
- //Fill the map of unigrams
- int lineno = 1;
- while (uniScanner.hasNext()) {
-
- String words = uniScanner.next();
- Double d = Double.valueOf(uniScanner.next());
- unimap.put(words.toLowerCase(), d);
- if (uniScanner.hasNextLine()) {
- uniScanner.nextLine();
- }
- lineno++;
-
- }
-
- //fill the map of bigrams
- while (biScanner.hasNext()) {
- String words = biScanner.next() + " " + biScanner.next();
- bimap.put(words.toLowerCase(), Double.valueOf(biScanner.next()));
- if (biScanner.hasNextLine()) {
- biScanner.nextLine();
- }
- }
- }
- System.err.println("Lexicons are read.");
- }
-
- /**
- * Executes a query that the analyzer can analyze.
- *
- * @param query The query string to execute.
- * @throws SQLException When database connection isn't available.
- */
- public void query(String query) throws SQLException {
- PreparedStatement statement;
- //make a connection to the database and execute the query
- statement = connection.prepareStatement(query);
- data = statement.executeQuery();
- }
-
- /**
- * Run a sentiment analysis and fill the database with the output.
- *
- * @param query The sql text for the query.
- * @throws SQLException
- * @throws IOException
- */
- public void sentimentAnalysis(String query) throws SQLException, IOException {
- query(query);
-
- //read the lexicons
- readLexicon();
-
- //go to the start of te dataset
- if (data == null) {
- System.err.println("data is empty, try querying first");
- return;
- }
-
- Double value;
- String text;
-
- //for all tuples
- while (data.next()) {
- //get the text
- text = data.getString("text");
- text = splitPunctToWords(text);
- // test is the tweet text you are going to analyze
- String[] words = text.split("\\s+"); // text splitted into separate words
- double positiverate = 0; // positive rating
-
- // Rate the text with unigrams
- for (String word : words) {
- value = unimap.get(word);
- if (value != null) {
- positiverate += unimap.get(word);
- }
- }
- // Rate the text with bigrams
- for (int i = 0; i < words.length - 1; i++) {
- String pair = words[i] + " " + words[i + 1];
- value = bimap.get(pair);
- if (value != null) {
- positiverate += bimap.get(pair);
- }
- }
- //insert the rating into the database
- NamedPreparedStatement m_insertRating;
- m_insertRating = new NamedPreparedStatement(connection, QueryUtils.insertRating);
- QueryUtils.setInsertParams(m_insertRating, data.getLong("tweetid"), data.getString("brand"), (int) (positiverate * 10));
- m_insertRating.executeUpdate();
- //don't print the rate
- //System.out.println(text + ": " + (int) (positiverate * 10));
- }
- }
-
- /**
- * Make a wordcloud of the results of some query.
- *
- * @param query The sql text for a query.
- * @throws SQLException
- * @throws FileNotFoundException
- * @throws UnsupportedEncodingException
- */
- public void makeWordCloud(String query) throws SQLException, FileNotFoundException, UnsupportedEncodingException {
-
- query(query);
- //go to the start of the ResultSet data
- if (data == null) {
- System.err.println("data is empty, try querying first");
- return;
- }
-
- String text;
- String[] words;
- HashMap<String, Integer> wordcloud = new HashMap<>();
-
- while (data.next()) {
- //get the text
- text = data.getString("text");
- //remove punctuation, convert to lowercase and split on words
- text = removePunct(text);
- text = text.toLowerCase();
- words = text.split("\\s+");
-
- for (String word : words) {
- if(wordcloud.containsKey(word)){
- wordcloud.put(word, wordcloud.get(word) + 1);
- }
- else{
- wordcloud.put(word, 1);
- }
- }
- }
- wordcloud.remove("");
- //print the words and their frequency in a csv file
- PrintWriter writer = new PrintWriter("wordcloud.csv", "UTF-8");
-
- writer.println("word,count");
-
- for(Entry e : wordcloud.entrySet()){
- writer.println(e.getKey() + "," + e.getValue());
- }
-
- writer.close();
- System.out.println("csv file made, please put it next to wordcloud.html and run this");
- }
-
- //generate csv for disco from the query
- public void disco(String query) throws SQLException, FileNotFoundException, UnsupportedEncodingException {
- //do the query
- query(query);
- PrintWriter writer = new PrintWriter("output.csv", "UTF-8");
- //print the first row
- for (int i = 1; i < data.getMetaData().getColumnCount(); i++) {
- writer.print(data.getMetaData().getColumnLabel(i) + ", ");
- }
- writer.println(data.getMetaData().getColumnLabel(data.getMetaData().getColumnCount()));
- //print the values
- while (data.next()) {
- for (int i = 1; i < data.getMetaData().getColumnCount(); i++) {
- if (data.getObject(i) == null) {
- writer.print(", ");
- } else {
- writer.print(data.getObject(i).toString().replaceAll("[,\n]", " ") + ", ");
- }
- }
- if (data.getObject(data.getMetaData().getColumnCount()) == null) {
- writer.println("0");
- } else {
- writer.println(data.getObject(data.getMetaData().getColumnCount()).toString().replace(",", " "));
- }
- }
- writer.close();
- }
-
- public void getBrands() throws SQLException {
- PreparedStatement statement;
- //make a connection to the database and execute the query
- statement = connection.prepareStatement("delete from mentionsbrand");
- statement.executeUpdate();
- BrandChecker checker = new BrandChecker("brandrules.txt");
- query("select * from tweet");
- NamedPreparedStatement m_insertBrand = new NamedPreparedStatement(connection, QueryUtils.insertBrand);
- while (data.next()) {
- List<String> brands = checker.getBrands(data.getString("text"));
- if (brands.isEmpty()) {
- QueryUtils.setInsertBrandParams(m_insertBrand, data.getLong("tweetid"), "no");
- m_insertBrand.executeUpdate();
- } else {
- for (String brand : brands) {
- QueryUtils.setInsertBrandParams(m_insertBrand, data.getLong("tweetid"), brand);
- m_insertBrand.executeUpdate();
- }
- }
-
- }
- }
-
- //replaces punctuation so it will be splitted
- //also removes urls
- private String splitPunctToWords(String text) {
- text = text.replaceAll("https?://\\S*", "");
- text = text.replaceAll("[!?):;\"']", " $0");
- text = text.replaceAll("[.,-](\\s|$)", " $0");
- text = text.replaceAll("\\s[(\"']", "$0 ");
- return text;
- }
-
- //removes punctuation
- //also removes urls
- private String removePunct(String text) {
- text = text.replaceAll("https?://\\S*", " ");
- text = text.replaceAll("@\\S*", " ");
- text = text.replaceAll("[^a-zA-Z0-9#_-]", " ");
- return text;
- }
-}
+package main; + +import analysis.BrandChecker; +import database.NamedPreparedStatement; +import database.QueryUtils; +import java.io.File; +import java.io.FileNotFoundException; +import java.io.IOException; +import java.io.PrintWriter; +import java.io.UnsupportedEncodingException; +import java.sql.Connection; +import java.sql.PreparedStatement; +import java.sql.ResultSet; +import java.sql.SQLException; +import java.util.List; +import java.util.HashMap; +import java.util.Map.Entry; +import java.util.Scanner; + +/** + * The sentiment analysis class that rates tweets based on a unigram and bigram + * set of weights. + */ +public class Analyzor { + + /** + * The map that matches single words to their weights. + */ + private final HashMap<String, Double> unimap = new HashMap(); + + /** + * The map that matches word pairs to their weights. + */ + private final HashMap<String, Double> bimap = new HashMap(); + + /** + * The results of a query, maybe return from query(). + */ + private ResultSet data; + + /** + * The persistent connection to the database. + */ + private final Connection connection; + + /** + * @param connection An open connection to the database. + */ + public Analyzor(Connection connection) { + this.connection = connection; + } + + /** + * Read the unigram and bigram lexica. + * + * @throws FileNotFoundException + */ + public void readLexicon() throws FileNotFoundException { + if (!unimap.isEmpty()) { + // data is already read. + return; + } + System.err.println("Trying to read lexicons..."); + // A unigram is in the format (WS = whitespace): + // word <WS> rating <WS> ??? <WS> ?? + // A bigram has an two WS-separated words instead of one. + try (Scanner uniScanner = new Scanner(new File("unigrams-pmilexicon.txt")); + Scanner biScanner = new Scanner(new File("bigrams-pmilexicon.txt"));) { + //Fill the map of unigrams + int lineno = 1; + while (uniScanner.hasNext()) { + + String words = uniScanner.next(); + Double d = Double.valueOf(uniScanner.next()); + unimap.put(words.toLowerCase(), d); + if (uniScanner.hasNextLine()) { + uniScanner.nextLine(); + } + lineno++; + + } + + //fill the map of bigrams + while (biScanner.hasNext()) { + String words = biScanner.next() + " " + biScanner.next(); + bimap.put(words.toLowerCase(), Double.valueOf(biScanner.next())); + if (biScanner.hasNextLine()) { + biScanner.nextLine(); + } + } + } + System.err.println("Lexicons are read."); + } + + /** + * Executes a query that the analyzer can analyze. + * + * @param query The query string to execute. + * @throws SQLException When database connection isn't available. + */ + public void query(String query) throws SQLException { + PreparedStatement statement; + //make a connection to the database and execute the query + statement = connection.prepareStatement(query); + data = statement.executeQuery(); + } + + /** + * Run a sentiment analysis and fill the database with the output. + * + * @param query The sql text for the query. + * @throws SQLException + * @throws IOException + */ + public void sentimentAnalysis(String query) throws SQLException, IOException { + query(query); + + //read the lexicons + readLexicon(); + + //go to the start of te dataset + if (data == null) { + System.err.println("data is empty, try querying first"); + return; + } + + Double value; + String text; + + //for all tuples + while (data.next()) { + //get the text + text = data.getString("text"); + text = splitPunctToWords(text); + // test is the tweet text you are going to analyze + String[] words = text.split("\\s+"); // text splitted into separate words + double positiverate = 0; // positive rating + + // Rate the text with unigrams + for (String word : words) { + value = unimap.get(word); + if (value != null) { + positiverate += unimap.get(word); + } + } + // Rate the text with bigrams + for (int i = 0; i < words.length - 1; i++) { + String pair = words[i] + " " + words[i + 1]; + value = bimap.get(pair); + if (value != null) { + positiverate += bimap.get(pair); + } + } + //insert the rating into the database + NamedPreparedStatement m_insertRating; + m_insertRating = new NamedPreparedStatement(connection, QueryUtils.insertRating); + QueryUtils.setInsertParams(m_insertRating, data.getLong("tweetid"), data.getString("brand"), (int) (positiverate * 10)); + m_insertRating.executeUpdate(); + //don't print the rate + //System.out.println(text + ": " + (int) (positiverate * 10)); + } + } + + /** + * Make a wordcloud of the results of some query. + * + * @param query The sql text for a query. + * @throws SQLException + * @throws FileNotFoundException + * @throws UnsupportedEncodingException + */ + public void makeWordCloud(String query) throws SQLException, FileNotFoundException, UnsupportedEncodingException { + + query(query); + //go to the start of the ResultSet data + if (data == null) { + System.err.println("data is empty, try querying first"); + return; + } + + String text; + String brand; + String[] words; + HashMap<String,HashMap<String, Integer>> wordcloud = new HashMap<>(); + + while (data.next()) { + //get brand + brand=data.getString("brand"); + //make hashmap for each brand + if(!wordcloud.containsKey(brand)){ + wordcloud.put(brand, new HashMap<String,Integer>()); + } + //get the text + text = data.getString("text"); + //remove punctuation, convert to lowercase and split on words + text = removePunct(text); + text = text.toLowerCase(); + words = text.split("\\s+"); + + for (String word : words) { + if(wordcloud.get(brand).containsKey(word)){ + wordcloud.get(brand).put(word, wordcloud.get(brand).get(word) + 1); + } + else{ + wordcloud.get(brand).put(word, 1); + } + } + } + wordcloud.remove(""); + //print the words and their frequency in a csv file + PrintWriter writer = new PrintWriter("wordcloud.csv", "UTF-8"); + + writer.println("brand,word,count"); + + //loop over brands + for(Entry en : wordcloud.entrySet()){ + //loop over words + for(Entry e : wordcloud.get(en.getKey()).entrySet()){ + writer.println(en.getKey() + "," + e.getKey() + "," + e.getValue()); + } + } + + writer.close(); + System.out.println("csv file made, please put it next to wordcloud.html and run this"); + } + + //generate csv for disco from the query + public void disco(String query) throws SQLException, FileNotFoundException, UnsupportedEncodingException { + //do the query + query(query); + PrintWriter writer = new PrintWriter("output.csv", "UTF-8"); + //print the first row + for (int i = 1; i < data.getMetaData().getColumnCount(); i++) { + writer.print(data.getMetaData().getColumnLabel(i) + ", "); + } + writer.println(data.getMetaData().getColumnLabel(data.getMetaData().getColumnCount())); + //print the values + while (data.next()) { + for (int i = 1; i < data.getMetaData().getColumnCount(); i++) { + if (data.getObject(i) == null) { + writer.print(", "); + } else { + writer.print(data.getObject(i).toString().replaceAll("[,\n]", " ") + ", "); + } + } + if (data.getObject(data.getMetaData().getColumnCount()) == null) { + writer.println("0"); + } else { + writer.println(data.getObject(data.getMetaData().getColumnCount()).toString().replace(",", " ")); + } + } + writer.close(); + } + + public void getBrands() throws SQLException { + PreparedStatement statement; + //make a connection to the database and execute the query + statement = connection.prepareStatement("delete from mentionsbrand"); + statement.executeUpdate(); + BrandChecker checker = new BrandChecker("brandonlyrules.txt"); + query("select * from tweet"); + NamedPreparedStatement m_insertBrand = new NamedPreparedStatement(connection, QueryUtils.insertBrand); + while (data.next()) { + List<String> brands = checker.getBrands(data.getString("text")); + if (brands.isEmpty()) { + QueryUtils.setInsertBrandParams(m_insertBrand, data.getLong("tweetid"), "no"); + m_insertBrand.executeUpdate(); + } else { + for (String brand : brands) { + QueryUtils.setInsertBrandParams(m_insertBrand, data.getLong("tweetid"), brand); + m_insertBrand.executeUpdate(); + } + } + + } + } + + //replaces punctuation so it will be splitted + //also removes urls + private String splitPunctToWords(String text) { + text = text.replaceAll("https?://\\S*", ""); + text = text.replaceAll("[!?):;\"']", " $0"); + text = text.replaceAll("[.,-](\\s|$)", " $0"); + text = text.replaceAll("\\s[(\"']", "$0 "); + return text; + } + + //removes punctuation + //also removes urls + private String removePunct(String text) { + text = text.replaceAll("https?://\\S*", " "); + text = text.replaceAll("@\\S*", " "); + text = text.replaceAll("[^a-zA-Z0-9#_-]", " "); + return text; + } +} |