3.1代码提交

src/test/java/com/qizhen/healsphere/DataWriteTest31.java

@@ -14,6 +14,10 @@ import org.apache.commons.lang3.StringUtils;
 import org.apache.poi.hssf.usermodel.HSSFRow;
 import org.apache.poi.hssf.usermodel.HSSFSheet;
 import org.apache.poi.hssf.usermodel.HSSFWorkbook;
+import org.apache.poi.ss.usermodel.Row;
+import org.apache.poi.ss.usermodel.Sheet;
+import org.apache.poi.ss.usermodel.Workbook;
+import org.apache.poi.xssf.usermodel.XSSFWorkbook;
 import org.junit.Test;
 import org.junit.runner.RunWith;
 import org.springframework.beans.factory.annotation.Autowired;
@@ -22,9 +26,7 @@ import org.springframework.context.annotation.ComponentScan;
 import org.springframework.test.context.junit4.SpringRunner;
 import org.springframework.util.CollectionUtils;
 
-import java.io.File;
-import java.io.FileOutputStream;
-import java.io.IOException;
+import java.io.*;
 import java.util.ArrayList;
 import java.util.List;
 import java.util.Map;
@@ -39,107 +41,8 @@ public class DataWriteTest31 {
     @Autowired
     EntityService entityService;
 
-    private static String diseaseStr = "风湿性心脏病" +
-            ",原发性高血压" +
-            ",急性心肌梗死" +
-            ",急性冠脉综合征" +
-            ",冠状动脉粥样硬化性心脏病" +
-            ",心力衰竭" +
-            ",心源性休克" +
-            ",酒精中毒" +
-            ",咽炎" +
-            ",急性扁桃体炎" +
-            ",变应性鼻炎" +
-            ",慢性鼻窦炎" +
-            ",鼻出血" +
-            ",急性牙髓炎" +
-            ",牙周病" +
-            ",溃疡性口炎" +
-            ",反流性食管炎" +
-            ",胃溃疡" +
-            ",慢性萎缩性胃炎" +
-            ",功能性消化不良" +
-            ",急性阑尾炎" +
-            ",克罗恩病" +
-            ",肠梗阻" +
-            ",肝硬化" +
-            ",急性胰腺炎" +
-            ",过敏性皮炎" +
-            ",湿疹" +
-            ",银屑病" +
-            ",带状疱疹" +
-            ",类风湿性关节炎" +
-            ",系统性红斑狼疮" +
-            ",慢性肾炎" +
-            ",肾病综合征" +
-            ",慢性肾衰竭" +
-            ",肾结石" +
-            ",输尿管结石" +
-            ",膀胱炎" +
-            ",前列腺增生" +
-            ",子宫内膜异位症" +
-            ",卵巢囊肿" +
-            ",痛经" +
-            ",女性更年期综合征" +
-            ",异位妊娠" +
-            ",妊娠剧吐" +
-            ",鼻咽恶性肿瘤" +
-            ",肝恶性肿瘤" +
-            ",乳房恶性肿瘤" +
-            ",宫颈恶性肿瘤" +
-            ",多发性骨髓瘤" +
-            ",急性白血病" +
-            ",血管瘤" +
-            ",子宫平滑肌瘤" +
-            ",新生儿黄疸" +
-            ",新生儿腹泻" +
-            ",缺铁性贫血" +
-            ",地中海贫血" +
-            ",血友病" +
-            ",甲状腺功能减退症" +
-            ",甲状腺功能亢进症" +
-            ",桥本甲状腺炎" +
-            ",2型糖尿病" +
-            ",糖尿病" +
-            ",甲状旁腺功能减退症" +
-            ",卵巢早衰" +
-            ",坏血病" +
-            ",高脂血症" +
-            ",高尿酸血症" +
-            ",精神分裂症" +
-            ",抑郁症" +
-            ",化脓性脑膜炎" +
-            ",帕金森病" +
-            ",癫痫" +
-            ",偏头痛" +
-            ",脑梗死" +
-            ",脑卒中" +
-            ",阻塞性睡眠呼吸暂停综合征" +
-            ",肺栓塞" +
-            ",慢性肺源性心脏病" +
-            ",急性上呼吸道感染" +
-            ",病毒性肺炎" +
-            ",社区获得性肺炎" +
-            ",支气管肺炎" +
-            ",肺炎" +
-            ",支气管哮喘" +
-            ",呼吸衰竭" +
-            ",肺部感染" +
-            ",急性泪腺炎" +
-            ",急性泪囊炎" +
-            ",巩膜炎" +
-            ",角膜炎" +
-            ",中耳炎" +
-            ",风湿性关节炎" +
-            ",颈椎病" +
-            ",颈肩综合征" +
-            ",坐骨神经痛" +
-            ",肩周炎" +
-            ",骨质疏松" +
-            ",锁骨骨折" +
-            ",肱骨骨折" +
-            ",肩关节脱位";
-
+    private static boolean selfCheck = false;
+    private static String urlExcelPath = "C:\\Users\\17664\\Desktop\\疾病.xlsx";
 
     @Test
     public void writeNeo4j() {
@@ -147,7 +50,7 @@ public class DataWriteTest31 {
         String propertyStr = "并发症\t的并发症有哪些疾病";
         String[] properties = propertyStr.split(",");
         workbook = new HSSFWorkbook();//这里也可以设置sheet的Name
-        String fileName = "";//
+        String fileName = "temp";//
         HSSFSheet sheet = workbook.createSheet(fileName);
         int rows = 0;
         for(String property:properties){
@@ -218,56 +121,24 @@ public class DataWriteTest31 {
         return entityService.create(createEntity);
     }
 
-    public static void main(String[] args) {
-        /*String disease = "咽炎";
-        String relation = "临床表现";
-        String answer = "**咽炎的临床表现为急性咽炎**^[1]^。";
-        String format = "你是专门处理医学领域文本的关系抽取专家。你将在指定的文本中抽取其中“"+disease+"的"+relation+"”。\n" +
-                "\n" +
-                "#要求\n" +
-                "1、抽取的结果将以JSON数组的形式呈现。每个抽取的“"+relation+"”高度简洁、高度概括，不要要描述性的文字，文字尽量保持在12个字符以内！\n" +
-                "\n" +
-                "#示例1\n" +
-                "以抽取“主要症状”为例\n" +
-                "文本：\n" +
-                "臭汗症的主要症状有全身或局部多汗且有臭味^[1][3][4]^。\n" +
-                "\n" +
-                "局部性臭汗症多发于大汗腺所在部位，如腋窝、腹股沟、足部、肛周、外阴、脐部及女性乳房下等处，以足部、腋部臭汗症最为多见。腋窝臭汗症俗称狐臭，是一种特殊的刺鼻臭味。足部臭汗症常与足部多汗伴发，有刺鼻的臭味^[1][4]^。\n" +
-                "\n" +
-                "全身性臭汗症为一种与种族有关的生理现象，也可见于卫生习惯不良者，服食某些食物（如葱、蒜、芥末）或某些药物（如麝香）后，在个别人中可产生臭汗^[4]^。\n" +
-                "\n" +
-                "输出：[\"全身或局部多汗且有臭味\",\"局部性臭汗症\",\"腋窝臭汗症\",\"足部臭汗症常\",\"全身性臭汗症\"]\n" +
-                "\n" +
-                "#示例2\n" +
-                "以抽取“是否传染病”为例\n" +
-                "文本：\n" +
-                "**是**^[2][4][6]^。\n" +
-                "\n" +
-                "输出：[\"是\"]\n\n"+
-                "2、没有可抽取的“"+relation+"”，则返回空json数组。\n" +
-                "\n" +
-                "本次抽取的文本如下：\n\n";
-
-
-        String zhiling = format+ answer;
-        System.out.println(zhiling);
-        System.out.println(BaidubceUtil.getChatResponse(zhiling, BaidubceUtil.getAccessToken()));*/
-
+    public static void main(String[] args) throws Exception {
 
         String accessToken = BaidubceUtil.getAccessToken();
-        String propertyStr = "分期\t的分期有哪些分期";
+        String propertyStr = "就诊科室\t有哪些“相关科室”或“就诊科室”或“所属科室”（有多个科室时，选相关性最高的，最多给我三个科室）？";
 
-        saveExel(propertyStr, accessToken,"分期3.1");
+        saveExel(propertyStr, accessToken,"就诊科室");
 
     }
 
     static HSSFWorkbook workbook;
-
+    static Long totalCount = 0l;
+    static Long successCount = 0l;
+    static Long failCount = 0l;
+    static Long unkonwCount = 0l;
     private static void saveExel(String propertyStr, String accessToken,String fileName) {
         String[] properties = propertyStr.split(",");
         workbook = new HSSFWorkbook();//这里也可以设置sheet的Name
         HSSFSheet sheet = workbook.createSheet(fileName);
-        ;//工作表
         int rows = 0;
         for(String property:properties){
             List<Knowlege> data = getData(accessToken, property);
@@ -283,7 +154,16 @@ public class DataWriteTest31 {
                     row.createCell(6).setCellValue(temp.getRefenrece() == null ? "" : temp.getRefenrece());
                 }
             }
+
+            String successRate = String.format("%.2f", Double.valueOf(successCount)/ Double.valueOf(totalCount));
+            String unkonwRate = String.format("%.2f", Double.valueOf(unkonwCount)/ Double.valueOf(totalCount));
+            sheet.getRow(rows-1).createCell(7).setCellValue(successRate);
+            sheet.getRow(rows-1).createCell(8).setCellValue(unkonwRate);
             save(fileName);
+            totalCount = 0l;
+            successCount = 0l;
+            failCount = 0l;
+            unkonwCount = 0l;
         }
     }
 
@@ -303,13 +183,18 @@ public class DataWriteTest31 {
     private static List<Knowlege> getData(String accessToken,String property) {
         List<Knowlege> list = new ArrayList<>();
         try {
-            BaidubceUtil baidubceUtil = new BaidubceUtil();
-
-            String[] diseases = diseaseStr.split(",");
             String appId= "3b615957-f9b1-4811-9dfc-13dcee6e0a37";
-            for (int i=0;i<diseases.length;i++) {
+                InputStream fis = new FileInputStream(urlExcelPath);
+                Workbook urlWorkbook = new XSSFWorkbook(fis);
+                Sheet urlSheet = urlWorkbook.getSheetAt(0);
+
+                for (int rowNum = 0; rowNum <= urlSheet.getLastRowNum(); rowNum++) {
                 try {
-                    String disease = diseases[i];
+                    Row row = urlSheet.getRow(rowNum);
+                    String disease = row.getCell(0).getStringCellValue();
+                    if(StringUtils.isEmpty(disease)){
+                        continue;
+                    }
                     disease = disease.trim();
                     String[] split = property.split("\t");
                     if(split.length<2){
@@ -317,14 +202,13 @@ public class DataWriteTest31 {
                     }
                     String quetionParty = split[1];
                     String relation = split[0];
-                    String question = disease + quetionParty + "？";
+                    String question = "“"+disease+"”" + quetionParty + "？";
 
                     Map<String, String> result = QizhenAssistant.getChatResponse(question, QizhenAssistant.getConversationId(appId),appId);
                     String answer = result.get("answer");
                     String references = result.get("references");
                     String defaultReferences = result.get("defaultReferences");
                     String chatResponse = "";
-                    System.out.println(answer);
                     if (!("failed".equals(answer) || answer.contains(QizhenAssistant.noAnswer))) {
                         String format = "你是专门处理医学领域文本的关系抽取专家。你将在指定的文本中抽取其中“"+disease+"的"+relation+"”。\n" +
                                 "\n" +
@@ -351,7 +235,7 @@ public class DataWriteTest31 {
 
                         String zhiling = format+ answer;
                         System.out.println(zhiling);
-                        chatResponse = baidubceUtil.getChatResponse(zhiling, accessToken);
+                        chatResponse = BaidubceUtil.getChatResponse(zhiling, accessToken);
                         chatResponse = filte(chatResponse);
                         JSONArray jsonResult = new JSONArray();
                         if(!StringUtils.isBlank(chatResponse)){
@@ -372,6 +256,30 @@ public class DataWriteTest31 {
                                             tempReferenceJA.add(referenceJO.getString("title"));
                                         }
                                     }
+                                    if(selfCheck) {
+                                        String llmQuestion = "你是一个资深的医学专家，请用“是”、“否”和“不确定”回答用户的问题回答。\n" +
+                                                "\n" +
+                                                "#要求\n" +
+                                                "1、只回答“是”、“否”和“不确定”，不要有额外的信息。\n" +
+                                                "\n" +
+                                                "#示例\n" +
+                                                "用户输入：急性上呼吸道感染的治疗药物是否包括“利巴韦林”？\n" +
+                                                "输出：是\n\n" +
+                                                "请回答：" + disease + "的" + relation + "是否包括“" + name + "”？";
+                                        String llmAnswer = BaidubceUtil.getChatResponse(llmQuestion, accessToken);
+                                        totalCount++;
+                                        temp.put("LLM-question", llmQuestion);
+                                        temp.put("LLM-answer", llmAnswer);
+                                        if (!StringUtils.isEmpty(llmAnswer) && llmAnswer.length() < 7) {
+                                            if (llmAnswer.contains("是")) {
+                                                successCount++;
+                                            } else if (llmAnswer.contains("否")) {
+                                                failCount++;
+                                            }
+                                        } else {
+                                            unkonwCount++;
+                                        }
+                                    }
                                     temp.put("reference", tempReferenceJA);
                                     if(tempReferenceJA.size()<1){
                                         temp.put("defaultReferences", defaultReferences);