1 |
|
---|
2 | import java.util.zip.*;
|
---|
3 | import java.io.BufferedOutputStream;
|
---|
4 | import java.io.BufferedReader;
|
---|
5 | import java.io.BufferedWriter;
|
---|
6 | import java.io.File;
|
---|
7 | import java.io.FileInputStream;
|
---|
8 | import java.io.FileNotFoundException;
|
---|
9 | import java.io.FileOutputStream;
|
---|
10 | import java.io.FileReader;
|
---|
11 | import java.io.IOException;
|
---|
12 | import java.io.InputStreamReader;
|
---|
13 | import java.io.OutputStreamWriter;
|
---|
14 | import java.net.Socket;
|
---|
15 | import java.nio.charset.Charset;
|
---|
16 | import java.nio.charset.StandardCharsets;
|
---|
17 | import java.nio.file.Files;
|
---|
18 | import java.nio.file.Paths;
|
---|
19 | import java.util.LinkedList;
|
---|
20 | import java.util.List;
|
---|
21 |
|
---|
22 |
|
---|
23 | public class CapiscoTextToTopics {
|
---|
24 |
|
---|
25 | private static final int BUFFER_SIZE = 4096;
|
---|
26 |
|
---|
27 | public static void main(String[] args) {
|
---|
28 |
|
---|
29 |
|
---|
30 | if ((args.length==0) || (args.length!=3) || (args[0].endsWith("help"))) {
|
---|
31 | System.err.println();
|
---|
32 | System.err.println("Usage: CapiscoTextToTopic 'server URL' 'server port' 'input documents path'");
|
---|
33 | System.err.println();
|
---|
34 | System.exit(-1);
|
---|
35 | }
|
---|
36 |
|
---|
37 | String serverName = args[0];
|
---|
38 | int port = Integer.parseInt(args[1]);
|
---|
39 | String documentspath = args[2];
|
---|
40 |
|
---|
41 | try {
|
---|
42 | System.out.println("Connecting to " + serverName
|
---|
43 | + " on port " + port);
|
---|
44 | Socket client = new Socket(serverName, port);
|
---|
45 | System.out.println("Just connected to "
|
---|
46 | + client.getRemoteSocketAddress());
|
---|
47 | BufferedReader in = new BufferedReader(new InputStreamReader(client.getInputStream()));
|
---|
48 | BufferedWriter out= new BufferedWriter(new OutputStreamWriter(client.getOutputStream()));
|
---|
49 | File documents = new File(documentspath);
|
---|
50 |
|
---|
51 | unzipAll(documents);
|
---|
52 | processFiles(documents, in, out);
|
---|
53 |
|
---|
54 | client.close();
|
---|
55 |
|
---|
56 | }
|
---|
57 | catch(IOException e) {
|
---|
58 | System.out.println("error: "+ e.getMessage());
|
---|
59 | e.printStackTrace();
|
---|
60 | }
|
---|
61 | }
|
---|
62 |
|
---|
63 | private static void unzipAll(File file) throws IOException {
|
---|
64 |
|
---|
65 | if (file.isDirectory())
|
---|
66 | {
|
---|
67 | for (File f : file.listFiles())
|
---|
68 | {
|
---|
69 | unzipAll(f);
|
---|
70 | }
|
---|
71 | }
|
---|
72 | else
|
---|
73 | {
|
---|
74 | String filename = file.getName().toLowerCase();
|
---|
75 | if(filename.endsWith(".zip"))
|
---|
76 | {
|
---|
77 | ZipInputStream zipIn = new ZipInputStream(new FileInputStream(file.getAbsolutePath()));
|
---|
78 | ZipEntry entry = zipIn.getNextEntry();
|
---|
79 | // iterates over entries in the zip file
|
---|
80 | while (entry != null)
|
---|
81 | {
|
---|
82 | String filePath = file.getParent() + File.separator + entry.getName();
|
---|
83 | if (!entry.isDirectory())
|
---|
84 | {
|
---|
85 | // if the entry is a file, extracts it
|
---|
86 | extractFile(zipIn, filePath);
|
---|
87 | }
|
---|
88 | else
|
---|
89 | {
|
---|
90 | // if the entry is a directory, make the directory
|
---|
91 | File dir = new File(filePath);
|
---|
92 | dir.mkdir();
|
---|
93 | }
|
---|
94 | zipIn.closeEntry();
|
---|
95 | entry = zipIn.getNextEntry();
|
---|
96 | }
|
---|
97 | zipIn.close();
|
---|
98 | }
|
---|
99 | }
|
---|
100 |
|
---|
101 |
|
---|
102 | }
|
---|
103 |
|
---|
104 |
|
---|
105 | static List<String> getTopics(BufferedReader input) throws IOException
|
---|
106 | {
|
---|
107 | List<String> topics = new LinkedList<String>();
|
---|
108 | String line ="";
|
---|
109 | while(!line.equals("--message end--") && !line.equals("0--message end--"))
|
---|
110 | {
|
---|
111 | line=input.readLine();
|
---|
112 | String[] words = line.split("\\|");
|
---|
113 | if(words.length>1)
|
---|
114 | {
|
---|
115 | topics.add(words[1]);
|
---|
116 | }
|
---|
117 | }
|
---|
118 | return topics;
|
---|
119 | }
|
---|
120 |
|
---|
121 | static String readFile(String path, Charset encoding)
|
---|
122 | throws IOException
|
---|
123 | {
|
---|
124 | byte[] encoded = Files.readAllBytes(Paths.get(path));
|
---|
125 | return new String(encoded, encoding);
|
---|
126 | }
|
---|
127 |
|
---|
128 |
|
---|
129 | private static void extractFile(ZipInputStream zipIn, String filePath) throws IOException {
|
---|
130 | BufferedOutputStream bos = new BufferedOutputStream(new FileOutputStream(filePath));
|
---|
131 | byte[] bytesIn = new byte[BUFFER_SIZE];
|
---|
132 | int read = 0;
|
---|
133 | while ((read = zipIn.read(bytesIn)) != -1) {
|
---|
134 | bos.write(bytesIn, 0, read);
|
---|
135 | }
|
---|
136 | bos.close();
|
---|
137 | }
|
---|
138 |
|
---|
139 | static void processFiles(File file, BufferedReader input, BufferedWriter output) throws IOException {
|
---|
140 |
|
---|
141 | if (!file.exists())
|
---|
142 | {
|
---|
143 | System.out.println(file + " does not exist.");
|
---|
144 | }
|
---|
145 | if (file.isDirectory())
|
---|
146 | {
|
---|
147 | for (File f : file.listFiles())
|
---|
148 | {
|
---|
149 | processFiles(f, input, output);
|
---|
150 | }
|
---|
151 | }
|
---|
152 | else
|
---|
153 | {
|
---|
154 | String filename = file.getName().toLowerCase();
|
---|
155 |
|
---|
156 | if (filename.endsWith(".htm") || filename.endsWith(".html") || filename.endsWith(".txt"))
|
---|
157 | {
|
---|
158 | List<String> WMtopics;
|
---|
159 | String topicsfilepath = file.getAbsolutePath()+".topics";
|
---|
160 | File topicsfile = new File(topicsfilepath);
|
---|
161 | topicsfile.createNewFile();
|
---|
162 | FileOutputStream fop = null;
|
---|
163 | fop = new FileOutputStream(topicsfile);
|
---|
164 |
|
---|
165 | String content = readFile(file.getAbsolutePath(), StandardCharsets.UTF_8);
|
---|
166 | output.write("topics xxxxxx\n ");
|
---|
167 | output.write(content + "\n");
|
---|
168 | output.write("xxxxxx \n");
|
---|
169 | output.flush();
|
---|
170 | System.out.println(file.getAbsolutePath() + " send for analysis.");
|
---|
171 | WMtopics = getTopics(input);
|
---|
172 | System.out.println(file.getAbsolutePath() + " has concepts received.");
|
---|
173 | for (int i = 0; i < WMtopics.size(); i++)
|
---|
174 | {
|
---|
175 | byte[] contentInBytes = (WMtopics.get(i)+"|").getBytes();
|
---|
176 | fop.write(contentInBytes);
|
---|
177 | fop.flush();
|
---|
178 | }
|
---|
179 | fop.close();
|
---|
180 | System.out.println(file.getAbsolutePath() + " processed.");
|
---|
181 |
|
---|
182 | }
|
---|
183 | else
|
---|
184 | {
|
---|
185 | System.out.println("Skipped " + filename);
|
---|
186 | }
|
---|
187 | }
|
---|
188 | }
|
---|
189 |
|
---|
190 | }
|
---|