百度内容审核
pom.xml
xml
<!--百度内容审核SDK-->
<dependency>
<groupId>com.baidu.aip</groupId>
<artifactId>java-sdk</artifactId>
<version>4.15.7</version>
</dependency>SDK默认使用slf4j-simple包进行日志输出，若需要使用自定义日志实现，可去除slf4j-simple依赖包，再额外添加相应的日志实现包即可（一个项目中有多个slf4j项目包可能会导致冲突）
maven去除slf4j-simple依赖包示例：
xml<!--百度内容审核SDK--> <dependency> <groupId>com.baidu.aip</groupId> <artifactId>java-sdk</artifactId> <version>4.15.7</version> <exclusions> <exclusion> <groupId>org.slf4j</groupId> <artifactId>slf4j-simple</artifactId> </exclusion> </exclusions> </dependency>
默认返回JSONObject类型(org.json.JSONObject)
java
AipContentCensor acc = new AipContentCensor(APP_ID,API_KEY,SECRET_KEY);
JSONObject jsonObject = acc.textCensorUserDefined("你好世界");
System.out.println(jsonObject.toString());
JSONObject jsonObject1 = acc.textCensorUserDefined("傻逼");
System.out.println(jsonObject1.toString(2));
if (jsonObject1.has("data"))
System.out.println(jsonObject1.get("data").toString());toString打印示例
{"conclusion":"合规","log_id":16931468069262180,"isHitMd5":false,"conclusionType":1}
{
"conclusion": "不合规",
"log_id": 16931468074792029,
"data": [{
"msg": "存在低俗辱骂不合规",
"conclusion": "不合规",
"hits": [{
"wordHitPositions": [{
"positions": [[
0,
1
]],
"label": "500200",
"keyword": "傻逼"
}],
"probability": 1,
"datasetName": "百度默认文本反作弊库",
"words": ["傻逼"],
"modelHitPositions": [[
0,
1,
1
]]
}],
"subType": 5,
"conclusionType": 2,
"type": 12
}],
"isHitMd5": false,
"conclusionType": 2
}
[{"msg":"存在低俗辱骂不合规","conclusion":"不合规","hits":[{"wordHitPositions":[{"positions":[[0,1]],"label":"500200","keyword":"傻逼"}],"probability":1,"datasetName":"百度默认文本反作弊库","words":["傻逼"],"modelHitPositions":[[0,1,1]]}],"subType":5,"conclusionType":2,"type":12}]注意
JSONObject类型无法直接作为添加了@ResponseBody注解控制层的返回值，不是Spring MVC中默认支持的数据类型，可以使用toMap()方法转化为Map类型javaMap<String, Object> map = jsonObject1.toMap(); System.out.println(map); System.out.println(map.get("data"));打印结果
{conclusion=不合规, log_id=16931474587780050, data=[{msg=存在低俗辱骂不合规, conclusion=不合规, hits=[{wordHitPositions=[{positions=[[0, 1]], label=500200, keyword=傻逼}], probability=1.0, datasetName=百度默认文本反作弊库, words=[傻逼], modelHitPositions=[[0, 1, 1.0]]}], subType=5, conclusionType=2, type=12}], isHitMd5=false, conclusionType=2} [{msg=存在低俗辱骂不合规, conclusion=不合规, hits=[{wordHitPositions=[{positions=[[0, 1]], label=500200, keyword=傻逼}], probability=1.0, datasetName=百度默认文本反作弊库, words=[傻逼], modelHitPositions=[[0, 1, 1.0]]}], subType=5, conclusionType=2, type=12}]
工具类代码
AipContentCensorClientConfig.java
java
import com.baidu.aip.contentcensor.AipContentCensor;
import lombok.Data;
import org.springframework.boot.context.properties.ConfigurationProperties;
import org.springframework.context.annotation.Bean;
import org.springframework.stereotype.Component;
@ConfigurationProperties(prefix = "ruoyi.censor")
@Component
@Data
public class AipContentCensorClientConfig {
/**
* 百度云审核的AppID
*/
private String AppID;
/**
* 百度云审核的Api Key
*/
private String API_Key;
/**
* 百度云审核的Secret Key
*/
private String Secret_Key;
@Bean
AipContentCensor commonTextCensorClient() {
/**
* 可以选择在客户端中添加参数，参考 https://ai.baidu.com/ai-doc/ANTIPORN/ik3h6xdze
* 如：
* // 可选：设置网络连接参数
* client.setConnectionTimeoutInMillis(2000);
* client.setSocketTimeoutInMillis(60000);
*
* // 可选：设置代理服务器地址, http和socket二选一，或者均不设置
* client.setHttpProxy("proxy_host", proxy_port); // 设置http代理
* client.setSocketProxy("proxy_host", proxy_port); // 设置socket代理
*
* // 可选：设置log4j日志输出格式，若不设置，则使用默认配置
* // 也可以直接通过jvm启动参数设置此环境变量
* System.setProperty("aip.log4j.conf", "path/to/your/log4j.properties");
*/
return new AipContentCensor(AppID, API_Key, Secret_Key);
}
}配置信息在application.yaml文件下
ContentWithCensorStateEnum.java
java
/**
* 内容审核状态
*/
public enum ContentWithCensorStateEnum {
/**
* 正常状态
*/
ADD,
/**
* 删除状态
*/
REMOVE,
/**
* Ai审核不通过
*/
CENSOR_FAIL,
/**
* Ai审核疑似不通过
*/
CENSOR_SUSPECT,
/**
* Ai审核错误
*/
CENSOR_ERROR,
/**
* 人工审核不通过
*/
BLOCK
}CensorResult.java
java
import com.ruoyi.common.enums.ContentWithCensorStateEnum;
import lombok.AllArgsConstructor;
import lombok.Data;
import lombok.NoArgsConstructor;
@Data
@AllArgsConstructor
@NoArgsConstructor
public class CensorResult {
/**
* 内容是否审核通过
*/
Boolean isPass;
/**
* 审核结果
*/
ContentWithCensorStateEnum contentWithCensorStateEnum;
/**
* 文字审核结果的Json字符串
*/
String textCensorJson;
/**
* 图片审核结果的Json字符串
*/
String imageCensorJson;
}AipContentCensorClientUtils.java
java
import com.baidu.aip.contentcensor.AipContentCensor;
import com.baidu.aip.contentcensor.EImgType;
import com.ruoyi.common.enums.ContentWithCensorStateEnum;
import com.ruoyi.jkqcyl.domain.CensorResult;
import lombok.Data;
import org.json.JSONObject;
import org.springframework.beans.factory.annotation.Autowired;
import org.springframework.stereotype.Component;
import java.io.ByteArrayOutputStream;
import java.io.IOException;
import java.io.InputStream;
@Component
@Data
public class AipContentCensorClientUtils {
/**
* 百度文本审核，识别审核结果的JSON KEY
*/
private static String CENSOR_CONCLUSION_TYPE_KEY = "conclusionType";
@Autowired
private AipContentCensor commonTextCensorClient;
/**
* 获取常规文本审核结果
*
* @param content 内容
* @return 百度内容审核JSON
*/
public CensorResult getCommonTextCensorResult(String content) {
//如果内容为空，则直接返回
if (content == null || content.isEmpty()) {
return getCensorResult(null);
}
try {
JSONObject response = commonTextCensorClient.textCensorUserDefined(content);
return getCensorResult(response);
} catch (Exception exception) {
System.out.println(exception);
return getCensorResult(null);
}
}
/**
* 获取图片审核结果
*
* @param imagePath 图片的本地路径或者Url
* @return 百度图片审核JSON
*/
public CensorResult getImageCensorResult(String imagePath) {
//如果内容为空，则直接返回
if (imagePath == null || imagePath.isEmpty()) {
return getCensorResult(null);
}
try {
JSONObject response = null;
if (imagePath.contains("http")){
response = commonTextCensorClient.imageCensorUserDefined(imagePath, EImgType.URL, null);
}else {
response = commonTextCensorClient.imageCensorUserDefined(imagePath, EImgType.FILE, null);
}
return getCensorResult(response);
} catch (Exception exception) {
System.out.println(exception);
return getCensorResult(null);
}
}
public CensorResult getImageCensorResult(InputStream inputStream){
//如果内容为空，则直接返回
if (inputStream == null) {
return getCensorResult(null);
}
byte[] imgData = getImgData(inputStream);
try {
JSONObject response = commonTextCensorClient.imageCensorUserDefined(imgData, null);
return getCensorResult(response);
} catch (Exception exception) {
System.out.println(exception);
return getCensorResult(null);
}
}
/**
* 获取审核结果
*
* @param clientJsonObject 百度审核的JSON字段
* @return 审核结果
*/
private CensorResult getCensorResult(JSONObject clientJsonObject) {
//获取代表审核结果的字段
//审核结果类型，可取值1.合规，2.不合规，3.疑似，4.审核失败
int conclusionType;
//如果是null就直接判定为失败
if (clientJsonObject == null) {
conclusionType = 4;
} else {
conclusionType = clientJsonObject.getInt(CENSOR_CONCLUSION_TYPE_KEY);
}
try {
ContentWithCensorStateEnum result;
switch (conclusionType) {
case 1:
//合规情况
result = ContentWithCensorStateEnum.ADD;
break;
case 2:
//不合规情况
result = ContentWithCensorStateEnum.CENSOR_FAIL;
break;
case 3:
//疑似不合规
result = ContentWithCensorStateEnum.CENSOR_SUSPECT;
break;
default:
//审核失败和其他情况，都归结到censor_error上去
result = ContentWithCensorStateEnum.CENSOR_ERROR;
break;
}
//过审要求：只能是合规情况
//解释：因为百度云控制台是可以调节不合规和疑似不合规的参数值的，因此这里只写合规情况就可以了
boolean isPass = result == ContentWithCensorStateEnum.ADD;
return new CensorResult(isPass, result, clientJsonObject != null ? clientJsonObject.toString() : null, null);
} catch (Exception exception) {
System.out.println(exception);
//如果出错，就直接返回true
return new CensorResult(true, null, null, null);
}
}
/**
* 处理图片的输入流
* @param inputStream 图片的输入流
* @return 返回字节数组
*/
private byte[] getImgData(InputStream inputStream){
ByteArrayOutputStream baos = new ByteArrayOutputStream();
byte[] buffer = new byte[1024];
byte[] imgData = null;
int len;
try{
while ((len = inputStream.read(buffer)) != -1) {
baos.write(buffer, 0, len);
}
}catch (IOException e) {
e.printStackTrace();
}
imgData = baos.toByteArray();
return imgData;
}
}