|
| 1 | +package com.tgb.ccl.http.common.util; |
| 2 | + |
| 3 | +import java.io.BufferedInputStream; |
| 4 | +import java.io.ByteArrayOutputStream; |
| 5 | +import java.io.File; |
| 6 | +import java.io.FileInputStream; |
| 7 | +import java.io.FileOutputStream; |
| 8 | +import java.io.IOException; |
| 9 | +import java.util.HashMap; |
| 10 | +import java.util.Map; |
| 11 | +import java.util.Map.Entry; |
| 12 | + |
| 13 | +import org.apache.http.Header; |
| 14 | +import org.apache.http.client.HttpClient; |
| 15 | + |
| 16 | +import com.tgb.ccl.http.common.HttpConfig; |
| 17 | +import com.tgb.ccl.http.common.HttpHeader; |
| 18 | +import com.tgb.ccl.http.common.Utils; |
| 19 | +import com.tgb.ccl.http.exception.HttpProcessException; |
| 20 | +import com.tgb.ccl.http.httpclient.HttpClientUtil; |
| 21 | +import com.tgb.ccl.http.httpclient.builder.HCB; |
| 22 | + |
| 23 | +/** |
| 24 | + * 识别验证码,自拼接http报文信息 |
| 25 | + * |
| 26 | + * @author arron |
| 27 | + * @date 2016年3月24日 上午9:44:35 |
| 28 | + * @version 1.0 |
| 29 | + */ |
| 30 | +public class OldOCR { |
| 31 | + |
| 32 | + /** |
| 33 | + * 接口说明: |
| 34 | + * https://github.com/AvensLab/OcrKing/blob/master/线上识别http接口说明.txt |
| 35 | + */ |
| 36 | + private static final String apiUrl = "http://lab.ocrking.com/ok.html"; |
| 37 | + private static final String apiKey = PropertiesUtil.getProperty("OCR.key"); |
| 38 | + private static final String boundary = "----------------------------OcrKing_Client_Aven_s_Lab"; |
| 39 | + private static final String end="\r\n--" + boundary + "--\r\n"; |
| 40 | + private static final Header[] headers = HttpHeader.custom() |
| 41 | + .accept("text/html, image/gif, image/jpeg, *; q=.2, */*; q=.2") |
| 42 | + .userAgent("Mozilla/5.0 (Windows NT 5.1; zh-CN; rv:1.9.1.3) Gecko/20100101 Firefox/8.0") |
| 43 | + .contentType("multipart/form-data; boundary="+boundary) |
| 44 | + .referer("http://lab.ocrking.com/?javaclient0.1)") |
| 45 | + .build(); |
| 46 | + private static final Map<String, Object> map = getParaMap(); |
| 47 | + private static HttpClient client =null; //=HCB.custom().proxy("127.0.0.1", 8888).build(); |
| 48 | + |
| 49 | + public static void enableCatch(){ |
| 50 | + client =HCB.custom().proxy("127.0.0.1", 8888).build(); |
| 51 | + } |
| 52 | + public static void unEnableCatch(){ |
| 53 | + client =null; |
| 54 | + } |
| 55 | + |
| 56 | + //获取固定参数 |
| 57 | + private static Map<String, Object> getParaMap(){ |
| 58 | + //加载所有参数 |
| 59 | + Map<String , Object> map = new HashMap<String, Object>(); |
| 60 | + map.put("service", "OcrKingForCaptcha"); |
| 61 | + map.put("language", "eng"); |
| 62 | + map.put("charset", "7");//7-数字大写小写,5-数字大写字母 |
| 63 | + map.put("type", "http://www.unknown.com"); |
| 64 | + map.put("apiKey", apiKey); |
| 65 | + return map; |
| 66 | + } |
| 67 | + |
| 68 | + |
| 69 | + /** |
| 70 | + * 识别本地校验码(英文:字母+大小写) |
| 71 | + * |
| 72 | + * @param imgFilePath 验证码地址 |
| 73 | + * @return |
| 74 | + */ |
| 75 | + public static String ocrCode(String filePath){ |
| 76 | + return ocrCode(filePath, 0); |
| 77 | + } |
| 78 | + /** |
| 79 | + * 识别本地校验码(英文:字母+大小写) |
| 80 | + * |
| 81 | + * @param imgFilePath 验证码地址 |
| 82 | + * @param limitCodeLen 验证码长度(如果结果与设定长度不一致,则返回获取失败的提示) |
| 83 | + * @return |
| 84 | + */ |
| 85 | + @SuppressWarnings("resource") |
| 86 | + public static String ocrCode(String imgFilePath, int limitCodeLen){ |
| 87 | + byte[] data = null; |
| 88 | + String fileName = imgFilePath.replaceAll("[^/]*/|[^\\\\]*\\\\", ""); |
| 89 | + |
| 90 | + StringBuffer strBuf = new StringBuffer(); |
| 91 | + for (Entry<String, Object> entry : map.entrySet()) { |
| 92 | + strBuf.append("\r\n").append("--").append(boundary).append("\r\n"); |
| 93 | + strBuf.append("Content-Disposition: form-data; name=\"" + entry.getKey() + "\"\r\n\r\n"); |
| 94 | + strBuf.append(entry.getValue()); |
| 95 | + } |
| 96 | + strBuf.append("\r\n").append("--").append(boundary).append("\r\n"); |
| 97 | + strBuf.append("Content-Disposition: form-data; name=\"ocrfile\"; filename=\"" + fileName + "\"\r\n"); |
| 98 | + strBuf.append("Content-Type:application/octet-stream\r\n\r\n"); |
| 99 | + |
| 100 | + //读取文件 |
| 101 | + File f = new File(imgFilePath); |
| 102 | + if(!f.exists()){ |
| 103 | + return "Error:文件不存在!"; |
| 104 | + } |
| 105 | + |
| 106 | + //内容长度=参数长度+文件长度+结尾字符串长度 |
| 107 | + ByteArrayOutputStream bos = new ByteArrayOutputStream(strBuf.length()+(int)f.length()+end.length()); |
| 108 | + try { |
| 109 | + bos.write(strBuf.toString().getBytes());//转化参数内容 |
| 110 | + BufferedInputStream in = new BufferedInputStream(new FileInputStream(f)); |
| 111 | + int buf_size = 1024; |
| 112 | + int len = 0; |
| 113 | + byte[] buf = new byte[buf_size]; |
| 114 | + while (-1 != (len = in.read(buf, 0, buf_size))) { |
| 115 | + bos.write(buf, 0, len); |
| 116 | + } |
| 117 | + bos.write(end.getBytes()); |
| 118 | + data= bos.toByteArray(); |
| 119 | + } catch (IOException e) { |
| 120 | + Utils.exception(e); |
| 121 | + } |
| 122 | + |
| 123 | + Map<String , Object> m = new HashMap<String, Object>(); |
| 124 | + m.put(Utils.ENTITY_BYTES, data); |
| 125 | + |
| 126 | + String html; |
| 127 | + try { |
| 128 | + html = HttpClientUtil.post(HttpConfig.custom().client(client).url(apiUrl).headers(headers).map(m)); |
| 129 | + //System.out.println(html); |
| 130 | + String[] results = StringUtil.regex("<Result>([^<]*)</Result>\\s*<Status>([^<]*)</Status>", html); |
| 131 | + if(results.length>0){ |
| 132 | + //System.out.println(results[0]); |
| 133 | + if(Boolean.parseBoolean(results[1])){ |
| 134 | + if(limitCodeLen<=0 || limitCodeLen==results[0].length()){//不判断长度或者长度一致时,直接返回 |
| 135 | + return results[0]; |
| 136 | + }else{ |
| 137 | + return "Error:获取失败! 原因:识别结果长度为:"+results[0].length()+"(期望长度:"+limitCodeLen+")"; |
| 138 | + } |
| 139 | + }else{ |
| 140 | + return "Error:获取失败! 原因:"+results[0]; |
| 141 | + } |
| 142 | + } |
| 143 | + } catch (HttpProcessException e) { |
| 144 | + Utils.exception(e); |
| 145 | + } |
| 146 | + |
| 147 | + return "Error:获取失败!"; |
| 148 | + } |
| 149 | + |
| 150 | + |
| 151 | + |
| 152 | + /** |
| 153 | + * 直接获取网络验证码(验证码不刷新) |
| 154 | + * |
| 155 | + * @param imgUrl 验证码地址 |
| 156 | + * @return |
| 157 | + */ |
| 158 | + public static String ocrCode4Net(String imgUrl){ |
| 159 | + return ocrCode4Net(imgUrl, 0); |
| 160 | + } |
| 161 | + /** |
| 162 | + * 直接获取网络验证码(验证码不刷新) |
| 163 | + * |
| 164 | + * @param imgUrl 验证码地址 |
| 165 | + * @param limitCodeLen 验证码长度 |
| 166 | + * @return |
| 167 | + */ |
| 168 | + public static String ocrCode4Net(String imgUrl, int limitCodeLen){ |
| 169 | + Map<String, Object> map = getParaMap(); |
| 170 | + map.put("url", imgUrl); |
| 171 | + |
| 172 | + Header[] headers = HttpHeader.custom().userAgent("Mozilla/5.0 (Windows NT 5.1; zh-CN; rv:1.9.1.3) Gecko/20100101 Firefox/8.0").build(); |
| 173 | + |
| 174 | + try { |
| 175 | + String html = HttpClientUtil.post(HttpConfig.custom().client(client).url(apiUrl).headers(headers).map(map)); |
| 176 | + //System.out.println(html); |
| 177 | + String[] results = StringUtil.regex("<Result>([^<]*)</Result>\\s*<Status>([^<]*)</Status>", html); |
| 178 | + if(results.length>0){ |
| 179 | + //System.out.println(results[0]); |
| 180 | + if(Boolean.parseBoolean(results[1])){ |
| 181 | + if(limitCodeLen<=0 || limitCodeLen==results[0].length()){//不判断长度或者长度一致时,直接返回 |
| 182 | + return results[0]; |
| 183 | + }else{ |
| 184 | + return "Error:获取失败! 原因:识别结果长度为:"+results[0].length()+"(期望长度:"+limitCodeLen+")"; |
| 185 | + } |
| 186 | + }else{ |
| 187 | + return "Error:获取失败! 原因:"+results[0]; |
| 188 | + } |
| 189 | + } |
| 190 | + } catch (HttpProcessException e) { |
| 191 | + Utils.exception(e); |
| 192 | + } |
| 193 | + |
| 194 | + return "Error:获取失败!"; |
| 195 | + } |
| 196 | + |
| 197 | + |
| 198 | + /** |
| 199 | + * 直接获取网络验证码(通过获取图片流,然后识别验证码) |
| 200 | + * |
| 201 | + * @param config HttpConfig对象(设置cookie) |
| 202 | + * @param savePath 图片保存的完整路径(值为null时,不保存),如:c:/1.png |
| 203 | + * @return |
| 204 | + */ |
| 205 | + public static String ocrCode4Net(HttpConfig config, String savePath){ |
| 206 | + return ocrCode4Net(config, savePath, 0); |
| 207 | + } |
| 208 | + /** |
| 209 | + * 直接获取网络验证码(通过获取图片流,然后识别验证码) |
| 210 | + * |
| 211 | + * @param config HttpConfig对象(设置cookie) |
| 212 | + * @param savePath 图片保存的完整路径(值为null时,不保存),如:c:/1.png |
| 213 | + * @param limitCodeLen 验证码长度 |
| 214 | + * @return |
| 215 | + */ |
| 216 | + @SuppressWarnings("resource") |
| 217 | + public static String ocrCode4Net(HttpConfig config, String savePath, int limitCodeLen){ |
| 218 | + if(savePath==null || savePath.equals("")){//如果不保存图片,则直接使用图片地址的方式获取验证码 |
| 219 | + return ocrCode4Net(config.url(), limitCodeLen); |
| 220 | + } |
| 221 | + |
| 222 | + byte[] data = null; |
| 223 | + |
| 224 | + StringBuffer strBuf = new StringBuffer(); |
| 225 | + for (Entry<String, Object> entry : map.entrySet()) { |
| 226 | + strBuf.append("\r\n").append("--").append(boundary).append("\r\n"); |
| 227 | + strBuf.append("Content-Disposition: form-data; name=\"" + entry.getKey() + "\"\r\n\r\n"); |
| 228 | + strBuf.append(entry.getValue()); |
| 229 | + } |
| 230 | + strBuf.append("\r\n").append("--").append(boundary).append("\r\n"); |
| 231 | + strBuf.append("Content-Disposition: form-data; name=\"ocrfile\"; filename=\"" + "aaa" + "\"\r\n"); |
| 232 | + strBuf.append("Content-Type:application/octet-stream\r\n\r\n"); |
| 233 | + |
| 234 | + //下载图片 |
| 235 | + ByteArrayOutputStream out = new ByteArrayOutputStream(); |
| 236 | + try { |
| 237 | + out = (ByteArrayOutputStream) HttpClientUtil.down(config.client(client).out(out)); |
| 238 | + //本地测试,可以保存一下图片,方便核验 |
| 239 | + FileOutputStream fos = new FileOutputStream(savePath); |
| 240 | + fos.write(out.toByteArray()); |
| 241 | + |
| 242 | + ByteArrayOutputStream bos = new ByteArrayOutputStream(out.size()+strBuf.length()+end.length()); |
| 243 | + bos.write(strBuf.toString().getBytes()); |
| 244 | + bos.write(out.toByteArray()); |
| 245 | + bos.write(end.getBytes()); |
| 246 | + data= bos.toByteArray(); |
| 247 | + } catch (HttpProcessException e) { |
| 248 | + Utils.exception(e); |
| 249 | + } catch (IOException e) { |
| 250 | + Utils.exception(e); |
| 251 | + } |
| 252 | + |
| 253 | + Map<String , Object> m = new HashMap<String, Object>(); |
| 254 | + m.put(Utils.ENTITY_BYTES, data); |
| 255 | + |
| 256 | + String html; |
| 257 | + try { |
| 258 | + html = HttpClientUtil.post(config.client(client).url(apiUrl).headers(headers).map(m)); |
| 259 | + //System.out.println(html); |
| 260 | + String[] results = StringUtil.regex("<Result>([^<]*)</Result>\\s*<Status>([^<]*)</Status>", html); |
| 261 | + if(results.length>0){ |
| 262 | + //System.out.println(results[0]); |
| 263 | + if(Boolean.parseBoolean(results[1])){ |
| 264 | + if(limitCodeLen<=0 || limitCodeLen==results[0].length()){//不判断长度或者长度一致时,直接返回 |
| 265 | + return results[0]; |
| 266 | + }else{ |
| 267 | + return "Error:获取失败! 原因:识别结果长度为:"+results[0].length()+"(期望长度:"+limitCodeLen+")"; |
| 268 | + } |
| 269 | + }else{ |
| 270 | + return "Error:获取失败! 原因:"+results[0]; |
| 271 | + } |
| 272 | + } |
| 273 | + } catch (HttpProcessException e) { |
| 274 | + Utils.exception(e); |
| 275 | + } |
| 276 | + |
| 277 | + return "Error:获取失败!"; |
| 278 | + } |
| 279 | + |
| 280 | + public static void main(String[] args) throws HttpProcessException, IOException { |
| 281 | +// enableCatch(); |
| 282 | + String filePath="C:/Users/160049/Desktop/中国.png"; |
| 283 | + String url = "http://file.ocrking.net:6080/small/20161104/w4fCjnzCl8KTwphpwqnCv2bCn8Kp/66fcff8d-61b1-49d6-bbfe-7428cf7accdf_debug.png?e9gFvJmkLbmgsZNTUCCNkjfi8J0Wbpn1CZHeP98eT1kxZ0ISBDt8Ql6h6zQ79pJg"; |
| 284 | + String url2 = "http://59.41.9.91/GZCX/WebUI/Content/Handler/ValidateCode.ashx?0.3271647585525703"; |
| 285 | + String code1 = ocrCode(filePath, 5); |
| 286 | + String code2 = ocrCode4Net(url,5); |
| 287 | + String code3 = ocrCode4Net(HttpConfig.custom().url(url2), System.getProperty("user.dir")+System.getProperty("file.separator")+"123.png", 5); |
| 288 | + System.out.println(code1); |
| 289 | + System.out.println(code2); |
| 290 | + System.out.println(code3); |
| 291 | + System.out.println("----"); |
| 292 | + } |
| 293 | +} |
0 commit comments