优化dify返回内容解析
This commit is contained in:
@@ -35,6 +35,11 @@ public class AuditContentRequest {
|
||||
*/
|
||||
private String history;
|
||||
|
||||
/**
|
||||
* 知识库关键词
|
||||
*/
|
||||
private String keywords;
|
||||
|
||||
/**
|
||||
* 优化建议
|
||||
*/
|
||||
|
||||
@@ -25,6 +25,8 @@ import org.springframework.beans.factory.annotation.Autowired;
|
||||
import java.util.*;
|
||||
import java.util.concurrent.CompletableFuture;
|
||||
import java.util.concurrent.ConcurrentHashMap;
|
||||
import java.util.regex.Matcher;
|
||||
import java.util.regex.Pattern;
|
||||
import java.util.stream.Collectors;
|
||||
|
||||
@Slf4j
|
||||
@@ -50,7 +52,7 @@ public abstract class AbstractAuditContentService {
|
||||
protected JSONArray callWorkflow(String url, String token, JSONObject requestBody, String workflowName) {
|
||||
try {
|
||||
log.info("调用{}工作流，请求体长度: {}", workflowName, requestBody.toString().length());
|
||||
|
||||
|
||||
String result = HttpUtil.createPost(url)
|
||||
.header("Authorization", token)
|
||||
.header("Content-Type", "application/json")
|
||||
@@ -58,34 +60,95 @@ public abstract class AbstractAuditContentService {
|
||||
.timeout(10 * 60 * 1000)
|
||||
.execute()
|
||||
.body();
|
||||
|
||||
|
||||
log.info("{}工作流返回结果长度: {}", workflowName, result.length());
|
||||
|
||||
|
||||
ObjectMapper objectMapper = new ObjectMapper();
|
||||
JsonNode rootNode = objectMapper.readTree(result);
|
||||
|
||||
|
||||
String outputText = rootNode.path("data")
|
||||
.path("outputs")
|
||||
.path("result")
|
||||
.asText();
|
||||
|
||||
.path("outputs")
|
||||
.path("result")
|
||||
.asText();
|
||||
|
||||
if (StrUtil.isBlank(outputText)) {
|
||||
log.warn("{}工作流返回结果为空", workflowName);
|
||||
log.warn("{}工作流返回 result 为空", workflowName);
|
||||
return new JSONArray();
|
||||
}
|
||||
|
||||
JsonNode arrayNode = objectMapper.readTree(outputText);
|
||||
JSONArray jsonArray = JSONArray.parseArray(arrayNode.toString());
|
||||
|
||||
|
||||
// ★ NEW：先从 ```json``` 中提取
|
||||
String jsonText = extractJsonFromMarkdown(outputText);
|
||||
|
||||
// ★ NEW：兜底提取
|
||||
if (StrUtil.isBlank(jsonText)) {
|
||||
jsonText = extractFirstJson(outputText);
|
||||
}
|
||||
|
||||
if (StrUtil.isBlank(jsonText)) {
|
||||
log.error("{} 工作流返回内容无法解析为 JSON，原始内容：{}", workflowName, outputText);
|
||||
throw new RuntimeException("Dify 返回内容中未找到有效 JSON");
|
||||
}
|
||||
|
||||
JSONArray jsonArray = JSONArray.parseArray(jsonText);
|
||||
|
||||
log.info("成功解析{}工作流返回数据，记录数: {}", workflowName, jsonArray.size());
|
||||
return jsonArray;
|
||||
|
||||
|
||||
} catch (Exception e) {
|
||||
log.error("调用{}工作流失败", workflowName, e);
|
||||
throw new RuntimeException("调用" + workflowName + "工作流失败: " + e.getMessage(), e);
|
||||
}
|
||||
}
|
||||
|
||||
/**
|
||||
* 从 ```json 代码块中提取 JSON
|
||||
*/
|
||||
private static String extractJsonFromMarkdown(String text) {
|
||||
if (StrUtil.isBlank(text)) {
|
||||
return null;
|
||||
}
|
||||
Pattern pattern = Pattern.compile("```json\\s*(.*?)\\s*```", Pattern.DOTALL);
|
||||
Matcher matcher = pattern.matcher(text);
|
||||
if (matcher.find()) {
|
||||
return matcher.group(1).trim();
|
||||
}
|
||||
return null;
|
||||
}
|
||||
|
||||
/**
|
||||
* 兜底：从混合文本中提取第一个完整 JSON（{} 或 []）
|
||||
*/
|
||||
private static String extractFirstJson(String text) {
|
||||
if (StrUtil.isBlank(text)) return null;
|
||||
|
||||
int start = -1;
|
||||
char open = 0, close = 0;
|
||||
|
||||
for (int i = 0; i < text.length(); i++) {
|
||||
char c = text.charAt(i);
|
||||
if (c == '{' || c == '[') {
|
||||
start = i;
|
||||
open = c;
|
||||
close = (c == '{') ? '}' : ']';
|
||||
break;
|
||||
}
|
||||
}
|
||||
if (start == -1) return null;
|
||||
|
||||
int count = 0;
|
||||
for (int i = start; i < text.length(); i++) {
|
||||
char c = text.charAt(i);
|
||||
if (c == open) count++;
|
||||
if (c == close) count--;
|
||||
if (count == 0) {
|
||||
return text.substring(start, i + 1);
|
||||
}
|
||||
}
|
||||
return null;
|
||||
}
|
||||
|
||||
|
||||
|
||||
/**
|
||||
* 构建工作流请求通用方法
|
||||
*/
|
||||
@@ -208,11 +271,22 @@ public abstract class AbstractAuditContentService {
|
||||
String docName = extractDocumentName(node);
|
||||
String docId = extractDocumentId(node);
|
||||
String fileUrl = fileUrlMap.get(docId);
|
||||
if(StrUtil.isBlank(fileUrl)) {
|
||||
fileUrl = extractDocumentUrl(node);
|
||||
}
|
||||
if(StrUtil.isBlank(fileUrl)) {
|
||||
fileUrl = "无链接";
|
||||
}
|
||||
|
||||
// 处理文件URL为空的情况
|
||||
String url = StrUtil.isNotBlank(fileUrl) ? fileUrl : "无";
|
||||
// String url = StrUtil.isNotBlank(fileUrl) ? fileUrl : "无";
|
||||
// 格式化结果
|
||||
String formattedText = String.format("《%s》【FileUrl:%s】%s", docName, url, text);
|
||||
// String formattedText = String.format("《%s》【FileUrl:%s】%s", docName, url, text);
|
||||
JSONObject json = new JSONObject();
|
||||
json.put("document_name", docName);
|
||||
json.put("file_url", fileUrl);
|
||||
json.put("content", text);
|
||||
String formattedText = json.toJSONString();
|
||||
results.add(formattedText);
|
||||
} catch (Exception e) {
|
||||
log.warn("处理知识库节点失败", e);
|
||||
@@ -259,6 +333,25 @@ public abstract class AbstractAuditContentService {
|
||||
return "相关文档";
|
||||
}
|
||||
|
||||
/**
|
||||
* 提取文档url通用方法
|
||||
*/
|
||||
protected String extractDocumentUrl(RetrieveResponseBodyDataNodes node) {
|
||||
try {
|
||||
Object metadataObj = node.getMetadata();
|
||||
if (metadataObj instanceof Map) {
|
||||
Map<?, ?> metadata = (Map<?, ?>) metadataObj;
|
||||
Object docIdObj = metadata.get("doc_url");
|
||||
if (docIdObj != null) {
|
||||
return docIdObj.toString();
|
||||
}
|
||||
}
|
||||
} catch (Exception e) {
|
||||
log.debug("提取文档名称失败", e);
|
||||
}
|
||||
return "相关文档";
|
||||
}
|
||||
|
||||
/**
|
||||
* 构建成功响应通用方法
|
||||
*/
|
||||
|
||||
Reference in New Issue
Block a user