[14] | 1 | import java.io.BufferedReader; |
---|
| 2 | import java.io.BufferedWriter; |
---|
| 3 | import java.io.FileNotFoundException; |
---|
| 4 | import java.io.FileReader; |
---|
| 5 | import java.io.FileWriter; |
---|
| 6 | import java.io.IOException; |
---|
| 7 | import java.io.InputStream; |
---|
| 8 | import java.io.InputStreamReader; |
---|
| 9 | import java.io.Reader; |
---|
| 10 | import java.util.regex.Matcher; |
---|
| 11 | import java.util.regex.Pattern; |
---|
| 12 | |
---|
| 13 | import org.apache.commons.httpclient.HttpClient; |
---|
| 14 | import org.apache.commons.httpclient.HttpException; |
---|
| 15 | import org.apache.commons.httpclient.NameValuePair; |
---|
| 16 | import org.apache.commons.httpclient.methods.PostMethod; |
---|
| 17 | |
---|
| 18 | |
---|
| 19 | public class GetTranslation { |
---|
| 20 | |
---|
| 21 | private static String translate(String langpair,String text) { |
---|
| 22 | |
---|
| 23 | /********************************************************** |
---|
| 24 | * 要這一行動起來, 你要有一些相依的library才行 |
---|
| 25 | * commons-logging |
---|
| 26 | * commons-codec |
---|
| 27 | */ |
---|
| 28 | HttpClient client = new HttpClient(); |
---|
| 29 | |
---|
| 30 | |
---|
| 31 | /*********************************************************** |
---|
| 32 | * 設定連線的站臺位置 |
---|
| 33 | * 請別在"站臺位置"加了多餘的protocol |
---|
| 34 | * ex. setHost("http://tinyurl.com", 80, "http"); |
---|
| 35 | * 不然他就噴UnknownHostException給你:P |
---|
| 36 | */ |
---|
| 37 | client.getHostConfiguration().setHost("www.google.com", 80, "http"); |
---|
| 38 | |
---|
| 39 | /*********************************************************** |
---|
| 40 | * 觀察了表單內容是用post方法, 而使用的script name為create.php |
---|
| 41 | */ |
---|
| 42 | PostMethod post = new PostMethod("/translate_t"); |
---|
| 43 | //post.addParameter(new NameValuePair("url", url)); |
---|
| 44 | post.addParameter("langpair", langpair); |
---|
| 45 | post.addParameter("text", text); |
---|
| 46 | |
---|
| 47 | String s = null; |
---|
| 48 | try { |
---|
| 49 | /*********************************************************** |
---|
| 50 | * 透過http client來執行post/get方法 |
---|
| 51 | */ |
---|
| 52 | client.executeMethod(post); |
---|
| 53 | |
---|
| 54 | /*********************************************************** |
---|
| 55 | * 再用您選用的方法傳回response body |
---|
| 56 | * 這裡偷懶了一下使用了getResponseBodyAsString() |
---|
| 57 | * log上會發出警告 |
---|
| 58 | * |
---|
| 59 | * 2006/5/18 上午 08:13:22 org.apache.commons.httpclient.HttpMethodBase getResponseBody |
---|
| 60 | * 警告: Going to buffer response body of large or unknown size. |
---|
| 61 | * Using getResponseBodyAsStream instead is recommended. |
---|
| 62 | */ |
---|
| 63 | InputStream in = post.getResponseBodyAsStream(); |
---|
| 64 | BufferedReader buf = new BufferedReader(new InputStreamReader(in)); |
---|
| 65 | Pattern p = Pattern.compile(".*id=\"*result_box\"*\\ *dir=\"*ltr\"*\\>"); |
---|
| 66 | Matcher m = null; |
---|
| 67 | |
---|
| 68 | String line; |
---|
| 69 | while ((line = buf.readLine()) != null) { |
---|
| 70 | m = p.matcher(line); |
---|
| 71 | //System.out.println(line); |
---|
| 72 | if (m.find()) { |
---|
| 73 | s = line.substring(m.end()); |
---|
| 74 | line = s; |
---|
| 75 | p = Pattern.compile("\\<\\ */\\ *div\\ *\\>"); |
---|
| 76 | m = p.matcher(line); |
---|
| 77 | if (m.find()) { |
---|
| 78 | s = line.substring(0, m.start()); |
---|
| 79 | } |
---|
| 80 | break; |
---|
| 81 | } |
---|
| 82 | } |
---|
| 83 | |
---|
| 84 | //s = GetTranslation.instostr(in); |
---|
| 85 | } catch (HttpException e) { |
---|
| 86 | e.printStackTrace(); |
---|
| 87 | } catch (IOException e) { |
---|
| 88 | e.printStackTrace(); |
---|
| 89 | } |
---|
| 90 | return s; |
---|
| 91 | } |
---|
| 92 | |
---|
| 93 | public static String instostr (InputStream in) throws IOException { |
---|
| 94 | StringBuffer out = new StringBuffer(); |
---|
| 95 | byte[] b = new byte[4096]; |
---|
| 96 | for (int n; (n = in.read(b)) != -1;) { |
---|
| 97 | out.append(new String(b, 0, n)); |
---|
| 98 | } |
---|
| 99 | return out.toString(); |
---|
| 100 | } |
---|
| 101 | |
---|
| 102 | /** @param filePath the name of the file to open. Not sure if it can accept URLs or just filenames. Path handling could be better, and buffer sizes are hardcoded |
---|
| 103 | */ |
---|
| 104 | public static void translateFile(String langpair, String inPath,String outPath) { |
---|
| 105 | BufferedReader reader=null; |
---|
| 106 | BufferedWriter writer=null; |
---|
| 107 | try { |
---|
| 108 | reader = new BufferedReader(new FileReader(inPath)); |
---|
| 109 | writer = new BufferedWriter(new FileWriter(outPath)); |
---|
| 110 | } catch (FileNotFoundException e) { |
---|
| 111 | e.printStackTrace(); |
---|
| 112 | } catch (IOException e) { |
---|
| 113 | e.printStackTrace(); |
---|
| 114 | } |
---|
| 115 | Pattern p = Pattern.compile("="); |
---|
| 116 | Matcher m = null; |
---|
| 117 | int count=0; |
---|
| 118 | String var=null, val=null; |
---|
| 119 | String line=null, lines="", ret=""; |
---|
| 120 | try { |
---|
| 121 | while ((line=reader.readLine())!=null) { |
---|
| 122 | |
---|
| 123 | if (line.contains("=")) { |
---|
| 124 | |
---|
| 125 | m = p.matcher(line); |
---|
| 126 | if (m.find()) { |
---|
| 127 | count++; |
---|
| 128 | var = line.substring(0,m.start()); |
---|
| 129 | val = line.substring(m.end()+1,line.length()-1); |
---|
| 130 | lines+=var+"="+val+"\n"; |
---|
| 131 | //System.out.println(var+"='"+GetTranslation.translate(langpair,val)+"'"); |
---|
| 132 | |
---|
| 133 | } |
---|
| 134 | } else { |
---|
| 135 | writer.write(line); |
---|
| 136 | writer.newLine(); |
---|
| 137 | } |
---|
| 138 | if (count==200) { |
---|
| 139 | ret = GetTranslation.translate(langpair,lines); |
---|
| 140 | ret = ret.replace("=", "=\""); |
---|
| 141 | ret = ret.replace("<br>", "\"\n"); |
---|
| 142 | System.out.println(ret); |
---|
| 143 | writer.write(ret); |
---|
| 144 | lines=""; |
---|
| 145 | count=0; |
---|
| 146 | Thread.sleep(5000); |
---|
| 147 | } |
---|
| 148 | } |
---|
| 149 | if (count>0) { |
---|
| 150 | ret=lines; |
---|
| 151 | ret = GetTranslation.translate(langpair,lines); |
---|
| 152 | ret = ret.replace("=", "=\""); |
---|
| 153 | ret = ret.replace("<br>", "\"\n"); |
---|
| 154 | System.out.println(ret); |
---|
| 155 | writer.write(ret); |
---|
| 156 | lines=""; |
---|
| 157 | Thread.sleep(5000); |
---|
| 158 | } |
---|
| 159 | reader.close(); |
---|
| 160 | writer.close(); |
---|
| 161 | } catch (IOException e) { |
---|
| 162 | e.printStackTrace(); |
---|
| 163 | } catch (InterruptedException e) { |
---|
| 164 | e.printStackTrace(); |
---|
| 165 | } |
---|
| 166 | |
---|
| 167 | } |
---|
| 168 | |
---|
| 169 | public static void main(String[] args) throws HttpException, IOException { |
---|
| 170 | |
---|
| 171 | if(null == args || args.length < 3) { |
---|
| 172 | System.out.println("Usage: java GetTranslation language_pair original_file translated_file"); |
---|
| 173 | System.out.println("Example: java GetTranslation \"en|zh-TW\" \"en_US\" \"zh_TW\""); |
---|
| 174 | //String ret = GetTranslation.translate("en|zh-TW","hello"); |
---|
| 175 | //System.out.println("Return Value: "+ret); |
---|
| 176 | System.exit(1); |
---|
| 177 | } |
---|
| 178 | //System.out.println(GetTranslation.translate("en|zh-TW","hello")); |
---|
| 179 | //GetTranslation.translateFile("en|zh-TW","en_US","zh_TW"); |
---|
| 180 | GetTranslation.translateFile(args[0],args[1],args[2]); |
---|
| 181 | |
---|
| 182 | //GetTranslation.translateFile("en|zh-TW","en_US","zh_TW"); |
---|
| 183 | //String ret = GetTranslation.translate("en|zh-TW","hello"); |
---|
| 184 | //System.out.println("Return Value: "+ret); |
---|
| 185 | |
---|
| 186 | } |
---|
| 187 | } |
---|