暂存
This commit is contained in:
@@ -11,7 +11,11 @@ import com.tencentcloudapi.ocr.v20181119.models.SmartStructuralOCRV2Response;
|
||||
import org.json.JSONArray;
|
||||
import org.json.JSONObject;
|
||||
|
||||
import java.io.File;
|
||||
import java.io.FileInputStream;
|
||||
import java.io.IOException;
|
||||
import java.util.ArrayList;
|
||||
import java.util.Base64;
|
||||
import java.util.List;
|
||||
|
||||
public class OCRUtils {
|
||||
@@ -20,7 +24,7 @@ public class OCRUtils {
|
||||
//API的SecretKey
|
||||
private static final String SECRET_KEY = "QjphKo8zkHZigT8j9PVtFPJyfIvO3d6V";
|
||||
|
||||
public static String pdfIdentifyText(String ImageUrl, Integer PageNumber) {
|
||||
public static String pdfIdentifyText(String imageBase64, Integer pageNumber) {
|
||||
try {
|
||||
// 实例化一个认证对象,入参需要传入腾讯云账户 SecretId 和 SecretKey,此处还需注意密钥对的保密
|
||||
// 代码泄露可能会导致 SecretId 和 SecretKey 泄露,并威胁账号下所有资源的安全性。以下代码示例仅供参考,建议采用更安全的方式来使用密钥,请参见:https://cloud.tencent.com/document/product/1278/85305
|
||||
@@ -36,12 +40,11 @@ public class OCRUtils {
|
||||
OcrClient client = new OcrClient(cred, "ap-beijing", clientProfile);
|
||||
// 实例化一个请求对象,每个接口都会对应一个request对象
|
||||
SmartStructuralOCRV2Request req = new SmartStructuralOCRV2Request();
|
||||
req.setImageUrl(ImageUrl);
|
||||
req.setImageBase64(imageBase64);
|
||||
req.setIsPdf(true);
|
||||
req.setPdfPageNumber(PageNumber.longValue());
|
||||
String[] itemNames1 = {"申请人", "统一社会信用代码", "法定代表人", "住所", "联系地址", "委托代理人"
|
||||
, "联系电话", "电子邮件", "被申请人", "居民身份证号码", "住所", "联系电话", "电子邮件", "仲裁请求"
|
||||
, "事实和理由"};
|
||||
req.setPdfPageNumber(pageNumber.longValue());
|
||||
String[] itemNames1 = {"申请人", "统一社会信用代码", "负责人", "住所", "联系地址"
|
||||
, "委托代理人", "联系电话", "电子邮件", "被申请人", "居民身份证号码", "仲裁请求", "事实和理由"};
|
||||
req.setItemNames(itemNames1);
|
||||
// 返回的resp是一个SmartStructuralOCRV2Response的实例,与请求对象对应
|
||||
SmartStructuralOCRV2Response resp = client.SmartStructuralOCRV2(req);
|
||||
@@ -78,18 +81,19 @@ public class OCRUtils {
|
||||
}
|
||||
return null;
|
||||
}
|
||||
|
||||
public static void main(String[] args) {
|
||||
String input = "申请人:招商银行,住所:上海嘉定区,联系地址:上海徐汇区,法定代表人:招商法人,委托代理人:黄海龙,被申请人:白贵勇,住所:陕西,联系地址:陕西,仲裁请求:仲裁委主任依据《2022年版仲裁规则》第三十一条第(四)项、第三十二条的规定指定白贵勇仲裁员为仲裁庭的仲裁员。本案由白贵勇仲裁员成立仲裁庭进行审理。,事实和理由:(一)申请人的仲裁请求及事实和理由,申请人:债务纠纷";
|
||||
String[] fields = input.split(",");
|
||||
JSONObject jsonObject = new JSONObject(); // 创建一个空的JSON对象
|
||||
for (String field : fields) {
|
||||
// 对于每个字段,再按冒号分隔符拆分出键和值
|
||||
String[] keyValue = field.split(":");
|
||||
if (keyValue.length == 2) { // 对于合法的键值对,将其添加到JSON对象中
|
||||
jsonObject.put(keyValue[0], keyValue[1]);
|
||||
}
|
||||
public static String pdfConvertBase64(String pathUrl){
|
||||
try {
|
||||
File file = new File(pathUrl);
|
||||
FileInputStream fileInputStream = new FileInputStream(file);
|
||||
byte[] fileBytes = new byte[(int) file.length()];
|
||||
fileInputStream.read(fileBytes);
|
||||
fileInputStream.close();
|
||||
// 将字节数组转换为Base64值
|
||||
return Base64.getEncoder().encodeToString(fileBytes);
|
||||
} catch (IOException e) {
|
||||
e.printStackTrace();
|
||||
}
|
||||
System.out.println(jsonObject.toString());
|
||||
return null;
|
||||
}
|
||||
|
||||
}
|
||||
|
||||
@@ -19,8 +19,7 @@ import org.json.JSONObject;
|
||||
|
||||
|
||||
import java.io.*;
|
||||
import java.util.ArrayList;
|
||||
import java.util.List;
|
||||
import java.util.*;
|
||||
|
||||
public class Tset {
|
||||
public static void main(String[] args) {
|
||||
@@ -29,10 +28,15 @@ public class Tset {
|
||||
//API的SecretKey
|
||||
final String SECRET_KEY = "QjphKo8zkHZigT8j9PVtFPJyfIvO3d6V";
|
||||
|
||||
String pdfFilePath = "http://121.40.189.20:8000/API/uploadPath/upload/1.pdf";
|
||||
// String pdfFilePath = "http://121.40.189.20:9000/API/uploadPath/upload/2023/11/14/ca2ca4697e5449ff9b9d23f95b221f58.pdf";
|
||||
|
||||
// String pdfFilePath = "http://121.40.189.20:9000/API/uploadPath/upload/ca2ca4697e5449ff9b9d23f95b221f58.pdf";
|
||||
// String pdfFilePath = "http://121.40.189.20:9000/API/uploadPath/upload/2023/11/14/ca2ca4697e5449ff9b9d23f95b221f58.pdf";
|
||||
String pdfFilePath ="D:/home/unzip/b92a7291-441e-4fad-95b4-7305f206eabd/仲裁材料/二、案件基本材料/2-1 仲裁申请书.pdf";
|
||||
try{
|
||||
// 读取文件内容到字节数组
|
||||
byte[] fileBytes = readFileToBytes(pdfFilePath);
|
||||
// 将字节数组转换为Base64值
|
||||
String base64String = encodeBytesToBase64(fileBytes);
|
||||
|
||||
// 实例化一个认证对象,入参需要传入腾讯云账户 SecretId 和 SecretKey,此处还需注意密钥对的保密
|
||||
// 代码泄露可能会导致 SecretId 和 SecretKey 泄露,并威胁账号下所有资源的安全性。以下代码示例仅供参考,建议采用更安全的方式来使用密钥,请参见:https://cloud.tencent.com/document/product/1278/85305
|
||||
// 密钥可前往官网控制台 https://console.cloud.tencent.com/cam/capi 进行获取
|
||||
@@ -47,22 +51,23 @@ public class Tset {
|
||||
OcrClient client = new OcrClient(cred, "ap-beijing", clientProfile);
|
||||
// 实例化一个请求对象,每个接口都会对应一个request对象
|
||||
SmartStructuralOCRV2Request req = new SmartStructuralOCRV2Request();
|
||||
req.setImageUrl(pdfFilePath);
|
||||
req.setImageBase64(base64String);
|
||||
req.setIsPdf(true);
|
||||
req.setPdfPageNumber(1L);
|
||||
String[] itemNames1 = {"申请人", "住所"};
|
||||
String[] itemNames1 = {"申请人", "统一社会信用代码", "负责人", "住所", "联系地址"
|
||||
, "委托代理人", "联系电话", "电子邮件", "被申请人", "居民身份证号码", "仲裁请求", "事实和理由"};
|
||||
req.setItemNames(itemNames1);
|
||||
// 返回的resp是一个SmartStructuralOCRV2Response的实例,与请求对象对应
|
||||
SmartStructuralOCRV2Response resp = client.SmartStructuralOCRV2(req);
|
||||
// 输出json格式的字符串回包
|
||||
System.out.println(SmartStructuralOCRV2Response.toJsonString(resp));
|
||||
|
||||
|
||||
//解析数据
|
||||
String s = SmartStructuralOCRV2Response.toJsonString(resp);
|
||||
// 解析JSON数据
|
||||
JSONObject jsonObject = new JSONObject(s);
|
||||
JSONArray structuralList = jsonObject.getJSONArray("StructuralList");
|
||||
// 遍历StructuralList中的Groups,获取Key对应的AutoName和Value对应的AutoContent
|
||||
List<String> strings = new ArrayList<>();
|
||||
StringBuilder stringBuilder = new StringBuilder(); // 创建一个StringBuilder对象
|
||||
for (int i = 0; i < structuralList.length(); i++) {
|
||||
JSONArray groups = structuralList.getJSONObject(i).getJSONArray("Groups");
|
||||
for (int j = 0; j < groups.length(); j++) {
|
||||
@@ -73,16 +78,53 @@ public class Tset {
|
||||
JSONObject value = line.getJSONObject("Value");
|
||||
String autoName = key.getString("AutoName");
|
||||
String autoContent = value.getString("AutoContent");
|
||||
String text = autoName +":"+ autoContent;
|
||||
strings.add(text);
|
||||
String text = autoName + ":" + autoContent;
|
||||
if (stringBuilder.length() > 0) {
|
||||
stringBuilder.append(","); // 在已有内容的情况下添加逗号分隔符
|
||||
}
|
||||
stringBuilder.append(text); // 拼接当前的字符串
|
||||
}
|
||||
}
|
||||
}
|
||||
System.out.println(strings);
|
||||
|
||||
} catch (TencentCloudSDKException e) {
|
||||
System.out.println("stringBuilder======"+stringBuilder.toString());
|
||||
// 将字符串按逗号分隔符切割
|
||||
String[] fields = stringBuilder.toString().split(",");
|
||||
Map<String, List<String>> map = new HashMap<>();
|
||||
for (String field : fields) {
|
||||
// 对于每个字段,再按冒号分隔符拆分出键和值
|
||||
String[] keyValue = field.split(":");
|
||||
if (keyValue.length == 2) { // 对于合法的键值对,将其添加到JSON对象中
|
||||
String key = keyValue[0];
|
||||
String value = keyValue[1];
|
||||
// 判断Map中是否已存在该键
|
||||
if (map.containsKey(key)) {
|
||||
// 如果已存在,获取该键对应的值,并将新的值添加到集合中
|
||||
List<String> values = map.get(key);
|
||||
values.add(value);
|
||||
} else {
|
||||
// 如果不存在,创建一个新的集合,并将值添加到集合中
|
||||
List<String> values = new ArrayList<>();
|
||||
values.add(value);
|
||||
map.put(key, values);
|
||||
}
|
||||
}
|
||||
}
|
||||
System.out.println(map);
|
||||
} catch (Exception e) {
|
||||
System.out.println(e.toString());
|
||||
}
|
||||
|
||||
}
|
||||
private static byte[] readFileToBytes(String filePath) throws IOException {
|
||||
File file = new File(filePath);
|
||||
FileInputStream fileInputStream = new FileInputStream(file);
|
||||
byte[] fileBytes = new byte[(int) file.length()];
|
||||
fileInputStream.read(fileBytes);
|
||||
fileInputStream.close();
|
||||
return fileBytes;
|
||||
}
|
||||
|
||||
private static String encodeBytesToBase64(byte[] bytes) {
|
||||
return Base64.getEncoder().encodeToString(bytes);
|
||||
}
|
||||
}
|
||||
Reference in New Issue
Block a user