修改表提取方式为程序快速提取

pom.xml

@@ -90,6 +90,16 @@
             <artifactId>neo4j-java-driver</artifactId>
             <version>5.15.0</version>
         </dependency>
+        <dependency>
+            <groupId>org.commonmark</groupId>
+            <artifactId>commonmark</artifactId>
+            <version>0.21.0</version>
+        </dependency>
+        <dependency>
+            <groupId>org.commonmark</groupId>
+            <artifactId>commonmark-ext-gfm-tables</artifactId>
+            <version>0.21.0</version>
+        </dependency>
     </dependencies>
     <build>
         <plugins>

src/main/java/com/supervision/pdfqaserver/dto/EREDTO.java

@@ -2,6 +2,7 @@ package com.supervision.pdfqaserver.dto;
 
 import cn.hutool.core.collection.CollUtil;
 import cn.hutool.core.lang.UUID;
+import cn.hutool.core.util.NumberUtil;
 import cn.hutool.core.util.RandomUtil;
 import cn.hutool.core.util.StrUtil;
 import com.alibaba.fastjson.JSONArray;
@@ -109,6 +110,46 @@ public class EREDTO {
         return eredto;
     }
 
+
+
+    public static EREDTO fromHeadAndRows(List<String> heads,List<List<String>> rows,String truncationId) {
+
+        EREDTO eredto = new EREDTO();
+        if (CollUtil.isEmpty(heads) || CollUtil.isEmpty(rows)){
+            return eredto;
+        }
+        List<EntityExtractionDTO> entities = new ArrayList<>();
+        for (List<String> row : rows) {
+            if (CollUtil.isEmpty(row)){
+                continue;
+            }
+            EntityExtractionDTO entityExtractionDTO = new EntityExtractionDTO();
+            entityExtractionDTO.setEntity("行");
+            // 避免表格行名重复
+            entityExtractionDTO.setName("行-" +  RandomUtil.randomString(UUID.randomUUID().toString(), 10));
+            entityExtractionDTO.setTruncationId(truncationId);
+            List<TruncationERAttributeDTO> truncationErAttributeDTOS = new ArrayList<>();
+            for (int i = 0; i < heads.size(); i++) {
+                String key = heads.get(i);
+                if (StrUtil.isBlank(key)){
+                    continue;
+                }
+                key = StrUtil.trim(key);
+                String value = i < row.size() ? row.get(i) : "";
+                if (StrUtil.isBlank(value)){
+                    continue;
+                }
+                value = StrUtil.trim(value);
+                TruncationERAttributeDTO truncationErAttributeDTO = new TruncationERAttributeDTO(key, value, NumberUtil.isNumber(value) ? "1" : "0");
+                truncationErAttributeDTOS.add(truncationErAttributeDTO);
+            }
+            entityExtractionDTO.setAttributes(truncationErAttributeDTOS);
+            entities.add(entityExtractionDTO);
+        }
+        return eredto;
+
+
+    }
     public static EREDTO fromTableJson(String json,String truncationId) {
 
         EREDTO eredto = new EREDTO();

src/main/java/com/supervision/pdfqaserver/service/TableVisitor.java

@@ -0,0 +1,98 @@
+package com.supervision.pdfqaserver.service;
+
+import org.commonmark.ext.gfm.tables.*;
+import org.commonmark.node.*;
+
+import java.util.ArrayList;
+import java.util.List;
+
+public class TableVisitor extends AbstractVisitor {
+    private boolean inHeader = false;
+    private boolean inBody = false;
+    private List<String> currentRow = null;
+
+    private List<String> headers = new ArrayList<>();
+
+    private final List<List<String>> rows = new ArrayList<>();
+
+    @Override
+    public void visit(CustomBlock customBlock) {
+        if (customBlock instanceof TableBlock) {
+            handleTableBlock((TableBlock) customBlock);
+        } else {
+            super.visit(customBlock);
+        }
+    }
+
+    @Override
+    public void visit(CustomNode customNode) {
+        if (customNode instanceof TableHead) {
+            handleTableHead((TableHead) customNode);
+        } else if (customNode instanceof TableBody) {
+            handleTableBody((TableBody) customNode);
+        } else if (customNode instanceof TableRow) {
+            handleTableRow((TableRow) customNode);
+        } else if (customNode instanceof TableCell) {
+            handleTableCell((TableCell) customNode);
+        } else {
+            super.visit(customNode);
+        }
+    }
+
+    private void handleTableBlock(TableBlock tableBlock) {
+        // 重置状态
+        inHeader = false;
+        inBody = false;
+        visitChildren(tableBlock);
+    }
+
+    private void handleTableHead(TableHead tableHead) {
+        inHeader = true;
+        visitChildren(tableHead);
+        inHeader = false;
+    }
+
+    private void handleTableBody(TableBody tableBody) {
+        inBody = true;
+        visitChildren(tableBody);
+        inBody = false;
+    }
+
+    private void handleTableRow(TableRow tableRow) {
+        currentRow = new ArrayList<>();
+        visitChildren(tableRow);
+
+        if (inHeader) {
+            this.headers = currentRow;
+        } else if (inBody) {
+            this.rows.add(currentRow);
+        }
+    }
+
+    private void handleTableCell(TableCell tableCell) {
+        if (currentRow != null) {
+            currentRow.add(getTextContent(tableCell));
+        }
+        visitChildren(tableCell);
+    }
+
+    private String getTextContent(Node node) {
+        StringBuilder sb = new StringBuilder();
+        Node child = node.getFirstChild();
+        while (child != null) {
+            if (child instanceof Text) {
+                sb.append(((Text) child).getLiteral());
+            }
+            child = child.getNext();
+        }
+        return sb.toString().trim();
+    }
+
+    public List<String> getTableHeaders() {
+        return headers;
+    }
+
+    public List<List<String>> getTableRows() {
+        return rows;
+    }
+}

src/main/java/com/supervision/pdfqaserver/service/impl/KnowledgeGraphServiceImpl.java

@@ -208,6 +208,7 @@ public class KnowledgeGraphServiceImpl implements KnowledgeGraphService {
                 log.info("意图元数据识别完成,切分文档id:{},耗时:{}毫秒", truncateDTO.getId(),interval.intervalMs("makeOutDomainMetadata"));
                 // 保存意图数据
                 intentSize ++;
+                index ++;
                 List<Intention> intentions = intentionService.batchSaveIfAbsent(intents, pdfInfo.getDomainCategoryId(), pdfId.toString());
                 for (Intention intention : intentions) {
                     List<DomainMetadataDTO> metadataDTOS = domainMetadataDTOS.stream()
@@ -215,7 +216,7 @@ public class KnowledgeGraphServiceImpl implements KnowledgeGraphService {
                     domainMetadataService.batchSaveOrUpdateMetadata(metadataDTOS,intention.getId(), pdfInfo.getDomainCategoryId());
                 }
             }catch (Exception e){
-                intentSize ++;
+                index ++;
                 log.error("切分文档id:{},意图识别失败", truncateDTO.getId(), e);
             }
 
@@ -284,15 +285,13 @@ public class KnowledgeGraphServiceImpl implements KnowledgeGraphService {
             try {
                 if (StrUtil.equals(truncateDTO.getLayoutType(), String.valueOf(LayoutTypeEnum.TABLE.getCode()))){
                     log.info("切分文档id:{},表格类型数据,不进行意图识别...", truncateDTO.getId());
-                    /*EREDTO eredto = conversionPipeline.doEre(truncateDTO, new ArrayList<>());
+                    EREDTO eredto = conversionPipeline.doEre(truncateDTO, new ArrayList<>());
                     if (null == eredto){
                         log.info("切分文档id:{},命名实体识别结果为空...", truncateDTO.getId());
                         continue;
                     }
                     this.saveERE(eredto, truncateDTO.getId());
                     eredtos.add(eredto);
-                    */
-                    continue;
                 }
 
                 timer.start("makeOutTruncationIntent");

src/main/java/com/supervision/pdfqaserver/service/impl/TripleConversionPipelineImpl.java

@@ -18,6 +18,10 @@ import edu.stanford.nlp.pipeline.CoreSentence;
 import edu.stanford.nlp.pipeline.StanfordCoreNLP;
 import lombok.RequiredArgsConstructor;
 import lombok.extern.slf4j.Slf4j;
+import org.commonmark.Extension;
+import org.commonmark.ext.gfm.tables.TablesExtension;
+import org.commonmark.node.Node;
+import org.commonmark.parser.Parser;
 import org.springframework.stereotype.Service;
 import java.util.*;
 import java.util.stream.Collectors;
@@ -216,7 +220,7 @@ public class TripleConversionPipelineImpl implements TripleConversionPipeline {
                 return doTextEre(truncateDTO);
             }
 
-            return doTableEre(truncateDTO);
+            return doTableEreFast(truncateDTO);
         }
         log.warn("doEre:错误的布局类型: {}", truncateDTO.getLayoutType());
         return null;
@@ -225,7 +229,7 @@ public class TripleConversionPipelineImpl implements TripleConversionPipeline {
     /**
      * 切分文档
      * 切分规则：
-     *       文本类型: 以单句为最小单元，最大字数现在这1000字以内。单句超过1000字取完成的单句。
+     *       文本类型: 以单句为最小单元，最大字数现在这maxTextLength字以内。单句超过maxTextLength字取完成的单句。
      *      表格类型: 以4行数据为最小单元。
      * @param documents 文档列表
      * @return
@@ -252,16 +256,14 @@ public class TripleConversionPipelineImpl implements TripleConversionPipeline {
         // 创建管道
         StanfordCoreNLP pipeline = new StanfordCoreNLP(props);
         List<TruncateDTO> truncateDTOS = new ArrayList<>();
-        StringBuilder truncateTextBuild = new StringBuilder(1500);
-        DocumentDTO documentDTOLast = null;
+        StringBuilder truncateTextBuild = new StringBuilder(minTextLength + maxTextLength);
         for (DocumentDTO documentDTO : documentDTOList) {
-            documentDTOLast = documentDTO;
             String content = documentDTO.getContent();
-            if (StrUtil.isEmpty(content)){
+            if (StrUtil.isEmpty(content)) {
                 continue;
             }
             Integer layoutType = documentDTO.getLayoutType();
-            if (LayoutTypeEnum.TEXT.getCode() == layoutType){
+            if (LayoutTypeEnum.TEXT.getCode() == layoutType) {
                 // 如果是文本类型的布局，进行合并
                 CoreDocument document = new CoreDocument(content);
                 // 分析文本
@@ -293,18 +295,20 @@ public class TripleConversionPipelineImpl implements TripleConversionPipeline {
                 // 如果是表格类型的布局，进行切分
                 // 出现表格后，如果truncateTextBuild不为空，单独作为一个片段
                 if (!truncateTextBuild.isEmpty()) {
-                    truncateDTOS.add(new TruncateDTO(documentDTO, truncateTextBuild.toString()));
+                    TruncateDTO truncateDTO = new TruncateDTO(documentDTO, truncateTextBuild.toString());
+                    truncateDTO.setLayoutType(String.valueOf(LayoutTypeEnum.TEXT.getCode()));//强制设置为文本类型
+                    truncateDTOS.add(truncateDTO);
                 }
                 // 提前抽取表名
                 TableTitleDTO tableTitleDTO = this.extractTableTitle(documentDTO.getTitle());
-                if (null != tableTitleDTO && StrUtil.isNotEmpty(tableTitleDTO.getTitle())){
+                if (null != tableTitleDTO && StrUtil.isNotEmpty(tableTitleDTO.getTitle())) {
                     documentDTO.setTitle(tableTitleDTO.getTitle());
-                }else {
+                } else {
                     // 生成一个默认的表
-                    documentDTO.setTitle("tableName-"+ RandomUtil.randomString(10));
+                    documentDTO.setTitle("tableName-" + RandomUtil.randomString(10));
                 }
                 List<String> tableRows = StrUtil.split(documentDTO.getContent(), "\n").stream().filter(StrUtil::isNotEmpty).collect(Collectors.toList());
-                if (tableRows.size()<5){
+                if (tableRows.size() < 5) {
                     TruncateDTO truncateDTO = new TruncateDTO(documentDTO);
                     truncateDTOS.add(truncateDTO);
                     continue;
@@ -312,7 +316,7 @@ public class TripleConversionPipelineImpl implements TripleConversionPipeline {
                 String tableTitle = tableRows.get(0);
                 // 标题分割符
                 String tableTitleSplit = tableRows.get(1);
-                List<String> noTitleRows = tableRows.subList(2,tableRows.size()-1);
+                List<String> noTitleRows = tableRows.subList(2, tableRows.size() - 1);
                 List<List<String>> rows = CollUtil.split(noTitleRows, 4);
                 for (List<String> row : rows) {
                     StringBuilder sb = new StringBuilder();
@@ -331,8 +335,10 @@ public class TripleConversionPipelineImpl implements TripleConversionPipeline {
                 log.info("sliceDocuments:错误的布局类型: {}", layoutType);
             }
         }
-        if (!truncateTextBuild.isEmpty() && null != documentDTOLast) {
-            truncateDTOS.add(new TruncateDTO(documentDTOLast, truncateTextBuild.toString()));
+        if (!truncateTextBuild.isEmpty() && null != CollUtil.getLast(documentDTOList)) {
+            TruncateDTO truncateDTO = new TruncateDTO(CollUtil.getLast(documentDTOList), truncateTextBuild.toString());
+            truncateDTO.setLayoutType(String.valueOf(LayoutTypeEnum.TEXT.getCode()));//强制设置为文本类型
+            truncateDTOS.add(truncateDTO);
         }
         return truncateDTOS;
     }
@@ -465,6 +471,33 @@ public class TripleConversionPipelineImpl implements TripleConversionPipeline {
         log.info("doTableEre响应结果:{}", response);
         EREDTO eredto = EREDTO.fromTableJson(response, truncateDTO.getId());
         // 手动设置表格标题
+        manualSetTableTitle(truncateDTO, eredto);
+        return eredto;
+    }
+
+    private EREDTO doTableEreFast(TruncateDTO truncateDTO){
+        log.info("doTableEreFast:开始进行表格实体关系抽取,内容:{}", truncateDTO.getContent());
+        if (StrUtil.isEmpty(truncateDTO.getContent())){
+            return null;
+        }
+        List<Extension> extensions = Arrays.asList(TablesExtension.create());
+        Parser parser = Parser.builder().extensions(extensions).build();
+
+        Node document = parser.parse(truncateDTO.getContent());
+        TableVisitor visitor = new TableVisitor();
+        document.accept(visitor);
+
+        List<String> tableHeaders = visitor.getTableHeaders();
+        List<List<String>> tableRows = visitor.getTableRows();
+
+        EREDTO eredto = EREDTO.fromHeadAndRows(tableHeaders, tableRows, truncateDTO.getId());
+        // 手动设置表格标题
+        manualSetTableTitle(truncateDTO, eredto);
+        return eredto;
+
+    }
+
+    private void manualSetTableTitle(TruncateDTO truncateDTO, EREDTO eredto) {
         EntityExtractionDTO titleEntity = new EntityExtractionDTO();
         titleEntity.setEntity("表");
         titleEntity.setTruncationId(truncateDTO.getId());
@@ -478,7 +511,6 @@ public class TripleConversionPipelineImpl implements TripleConversionPipeline {
         }
         eredto.getEntities().add(titleEntity);
         eredto.setRelations(relations);
-        return eredto;
     }
 
     /**

src/test/java/com/supervision/pdfqaserver/PdfQaServerApplicationTests.java

@@ -17,6 +17,7 @@ import java.util.List;
 import java.util.Map;
 import java.util.stream.Collectors;
 import static org.neo4j.driver.Values.parameters;
+import org.commonmark.node.*;
 
 @Slf4j
 @SpringBootTest
@@ -153,13 +154,13 @@ class PdfQaServerApplicationTests {
     @Test
     public void metaDataTrainTest() {
 
-        knowledgeGraphService.metaDataTrain(13);
+        knowledgeGraphService.metaDataTrain(15);
     }
 
     @Test
     void generateGraphBaseTrainTest() {
 
-        knowledgeGraphService.generateGraphBaseTrain(13);
+        knowledgeGraphService.generateGraphBaseTrain(14);
     }