feat: 完成S02桥接能力并接通M09工作流文档解析闭环

- 新增统一文档解析桥接子域，封装 easy-agents 文档解析门面 - 支持工作流开始节点文件上传与素材选择的单文件对象输入 - DocNode 改为文档解析节点，PDF 走统一解析，非 PDF 保持默认读取
2026-04-14 19:57:32 +08:00
parent 855e93ecbf
commit a41b50959e
30 changed files with 2475 additions and 20 deletions
--- a/easyflow-modules/easyflow-module-ai/pom.xml
+++ b/easyflow-modules/easyflow-module-ai/pom.xml
@@ -41,6 +41,10 @@
            <groupId>com.easyagents</groupId>
            <artifactId>easy-agents-support</artifactId>
        </dependency>
+        <dependency>
+            <groupId>com.easyagents</groupId>
+            <artifactId>easy-agents-document-core</artifactId>
+        </dependency>
        <dependency>
            <groupId>com.easyagents</groupId>
            <artifactId>easy-agents-rag-retrieval</artifactId>
--- a/easyflow-modules/easyflow-module-ai/src/main/java/tech/easyflow/ai/document/exception/DocumentParseBridgeException.java
+++ b/easyflow-modules/easyflow-module-ai/src/main/java/tech/easyflow/ai/document/exception/DocumentParseBridgeException.java
@@ -0,0 +1,81 @@
+package tech.easyflow.ai.document.exception;
+
+/**
+ * 文档解析桥接层异常。
+ *
+ * <p>桥接层负责把底层文档服务异常和文件加载异常转换为稳定语义，
+ * 供上层业务按场景进行处理。</p>
+ *
+ * @author Codex
+ * @since 2026-04-14
+ */
+public class DocumentParseBridgeException extends RuntimeException {
+
+    private final String code;
+
+    public DocumentParseBridgeException(String code, String message) {
+        super(message);
+        this.code = code;
+    }
+
+    public DocumentParseBridgeException(String code, String message, Throwable cause) {
+        super(message, cause);
+        this.code = code;
+    }
+
+    /**
+     * 获取稳定错误码。
+     *
+     * @return 错误码
+     */
+    public String getCode() {
+        return code;
+    }
+
+    public static DocumentParseBridgeException serviceNotEnabled() {
+        return new DocumentParseBridgeException(
+            "service_not_enabled",
+            "统一文档解析服务未启用，请先配置 easy-agents.document.pdf.provider"
+        );
+    }
+
+    public static DocumentParseBridgeException unsupportedSource(String message) {
+        return new DocumentParseBridgeException("unsupported_source", message);
+    }
+
+    public static DocumentParseBridgeException sourceLoadFailed(String message, Throwable cause) {
+        return new DocumentParseBridgeException("source_load_failed", message, cause);
+    }
+
+    public static DocumentParseBridgeException requestBuildFailed(String message, Throwable cause) {
+        return new DocumentParseBridgeException("request_build_failed", message, cause);
+    }
+
+    public static DocumentParseBridgeException requestBuildFailed(String message) {
+        return new DocumentParseBridgeException("request_build_failed", message);
+    }
+
+    public static DocumentParseBridgeException parseFailed(String message, Throwable cause) {
+        return new DocumentParseBridgeException("parse_failed", message, cause);
+    }
+
+    public static DocumentParseBridgeException parseFailed(String message) {
+        return new DocumentParseBridgeException("parse_failed", message);
+    }
+
+    public static DocumentParseBridgeException taskFailed(String message, Throwable cause) {
+        return new DocumentParseBridgeException("task_failed", message, cause);
+    }
+
+    public static DocumentParseBridgeException taskFailed(String message) {
+        return new DocumentParseBridgeException("task_failed", message);
+    }
+
+    public static DocumentParseBridgeException resultFetchFailed(String message, Throwable cause) {
+        return new DocumentParseBridgeException("result_fetch_failed", message, cause);
+    }
+
+    public static DocumentParseBridgeException resultFetchFailed(String message) {
+        return new DocumentParseBridgeException("result_fetch_failed", message);
+    }
+}
--- a/easyflow-modules/easyflow-module-ai/src/main/java/tech/easyflow/ai/document/model/DocumentParseArtifacts.java
+++ b/easyflow-modules/easyflow-module-ai/src/main/java/tech/easyflow/ai/document/model/DocumentParseArtifacts.java
@@ -0,0 +1,63 @@
+package tech.easyflow.ai.document.model;
+
+import java.util.LinkedHashMap;
+import java.util.Map;
+
+/**
+ * 文档解析工件。
+ *
+ * @author Codex
+ * @since 2026-04-14
+ */
+public class DocumentParseArtifacts {
+
+    private Object middleJson;
+    private Object contentList;
+    private Object modelOutput;
+    private Map<String, Object> extraJsonArtifacts = new LinkedHashMap<String, Object>();
+    private Map<String, byte[]> extraBinaryArtifacts = new LinkedHashMap<String, byte[]>();
+
+    public Object getMiddleJson() {
+        return middleJson;
+    }
+
+    public void setMiddleJson(Object middleJson) {
+        this.middleJson = middleJson;
+    }
+
+    public Object getContentList() {
+        return contentList;
+    }
+
+    public void setContentList(Object contentList) {
+        this.contentList = contentList;
+    }
+
+    public Object getModelOutput() {
+        return modelOutput;
+    }
+
+    public void setModelOutput(Object modelOutput) {
+        this.modelOutput = modelOutput;
+    }
+
+    public Map<String, Object> getExtraJsonArtifacts() {
+        return extraJsonArtifacts;
+    }
+
+    public void setExtraJsonArtifacts(Map<String, Object> extraJsonArtifacts) {
+        this.extraJsonArtifacts = extraJsonArtifacts == null
+            ? new LinkedHashMap<String, Object>()
+            : extraJsonArtifacts;
+    }
+
+    public Map<String, byte[]> getExtraBinaryArtifacts() {
+        return extraBinaryArtifacts;
+    }
+
+    public void setExtraBinaryArtifacts(Map<String, byte[]> extraBinaryArtifacts) {
+        this.extraBinaryArtifacts = extraBinaryArtifacts == null
+            ? new LinkedHashMap<String, byte[]>()
+            : extraBinaryArtifacts;
+    }
+}
--- a/easyflow-modules/easyflow-module-ai/src/main/java/tech/easyflow/ai/document/model/DocumentParseScenario.java
+++ b/easyflow-modules/easyflow-module-ai/src/main/java/tech/easyflow/ai/document/model/DocumentParseScenario.java
@@ -0,0 +1,28 @@
+package tech.easyflow.ai.document.model;
+
+/**
+ * 文档解析场景预设。
+ *
+ * <p>场景由 easyflow 业务层传入，桥接层负责将场景映射为底层解析请求参数，
+ * 避免业务模块直接感知多个布尔开关。</p>
+ *
+ * @author Codex
+ * @since 2026-04-14
+ */
+public enum DocumentParseScenario {
+
+    /**
+     * 工作流文本提取场景，仅要求尽快返回可直接消费的文本结果。
+     */
+    WORKFLOW_TEXT,
+
+    /**
+     * 知识库导入场景，需要保留更多结构化工件供后续分块分析使用。
+     */
+    KNOWLEDGE_IMPORT,
+
+    /**
+     * 尽可能返回完整工件，供后续高级消费场景使用。
+     */
+    FULL_ARTIFACTS
+}
--- a/easyflow-modules/easyflow-module-ai/src/main/java/tech/easyflow/ai/document/model/DocumentParseTaskInfo.java
+++ b/easyflow-modules/easyflow-module-ai/src/main/java/tech/easyflow/ai/document/model/DocumentParseTaskInfo.java
@@ -0,0 +1,33 @@
+package tech.easyflow.ai.document.model;
+
+/**
+ * 任务聚合查询结果。
+ *
+ * <p>该对象在任务状态基础上按需附带最终解析结果。
+ * 当任务尚未完成时仅返回状态信息；当任务已完成时可同时返回标准化结果。</p>
+ *
+ * @author Codex
+ * @since 2026-04-14
+ */
+public class DocumentParseTaskInfo extends DocumentParseTaskStatus {
+
+    private DocumentParsedResult result;
+
+    /**
+     * 获取标准化解析结果。
+     *
+     * @return 解析结果；任务未完成时可能为空
+     */
+    public DocumentParsedResult getResult() {
+        return result;
+    }
+
+    /**
+     * 设置标准化解析结果。
+     *
+     * @param result 解析结果
+     */
+    public void setResult(DocumentParsedResult result) {
+        this.result = result;
+    }
+}
--- a/easyflow-modules/easyflow-module-ai/src/main/java/tech/easyflow/ai/document/model/DocumentParseTaskStatus.java
+++ b/easyflow-modules/easyflow-module-ai/src/main/java/tech/easyflow/ai/document/model/DocumentParseTaskStatus.java
@@ -0,0 +1,113 @@
+package tech.easyflow.ai.document.model;
+
+import java.util.ArrayList;
+import java.util.List;
+
+/**
+ * 标准化异步解析任务状态。
+ *
+ * @author Codex
+ * @since 2026-04-14
+ */
+public class DocumentParseTaskStatus {
+
+    private String taskId;
+    private String status;
+    private String backend;
+    private List<String> fileNames = new ArrayList<String>();
+    private String createdAt;
+    private String startedAt;
+    private String completedAt;
+    private String error;
+    private String statusUrl;
+    private String resultUrl;
+    private Integer queuedAhead;
+
+    public String getTaskId() {
+        return taskId;
+    }
+
+    public void setTaskId(String taskId) {
+        this.taskId = taskId;
+    }
+
+    public String getStatus() {
+        return status;
+    }
+
+    public void setStatus(String status) {
+        this.status = status;
+    }
+
+    public String getBackend() {
+        return backend;
+    }
+
+    public void setBackend(String backend) {
+        this.backend = backend;
+    }
+
+    public List<String> getFileNames() {
+        return fileNames;
+    }
+
+    public void setFileNames(List<String> fileNames) {
+        this.fileNames = fileNames == null ? new ArrayList<String>() : fileNames;
+    }
+
+    public String getCreatedAt() {
+        return createdAt;
+    }
+
+    public void setCreatedAt(String createdAt) {
+        this.createdAt = createdAt;
+    }
+
+    public String getStartedAt() {
+        return startedAt;
+    }
+
+    public void setStartedAt(String startedAt) {
+        this.startedAt = startedAt;
+    }
+
+    public String getCompletedAt() {
+        return completedAt;
+    }
+
+    public void setCompletedAt(String completedAt) {
+        this.completedAt = completedAt;
+    }
+
+    public String getError() {
+        return error;
+    }
+
+    public void setError(String error) {
+        this.error = error;
+    }
+
+    public String getStatusUrl() {
+        return statusUrl;
+    }
+
+    public void setStatusUrl(String statusUrl) {
+        this.statusUrl = statusUrl;
+    }
+
+    public String getResultUrl() {
+        return resultUrl;
+    }
+
+    public void setResultUrl(String resultUrl) {
+        this.resultUrl = resultUrl;
+    }
+
+    public Integer getQueuedAhead() {
+        return queuedAhead;
+    }
+
+    public void setQueuedAhead(Integer queuedAhead) {
+        this.queuedAhead = queuedAhead;
+    }
+}
--- a/easyflow-modules/easyflow-module-ai/src/main/java/tech/easyflow/ai/document/model/DocumentParsedResult.java
+++ b/easyflow-modules/easyflow-module-ai/src/main/java/tech/easyflow/ai/document/model/DocumentParsedResult.java
@@ -0,0 +1,120 @@
+package tech.easyflow.ai.document.model;
+
+import com.easyagents.document.core.model.DocumentBlock;
+import com.easyagents.document.core.model.DocumentImage;
+import com.easyagents.document.core.model.DocumentPage;
+import com.easyagents.document.core.model.DocumentTable;
+
+import java.util.ArrayList;
+import java.util.LinkedHashMap;
+import java.util.List;
+import java.util.Map;
+
+/**
+ * 标准化单文档解析结果。
+ *
+ * @author Codex
+ * @since 2026-04-14
+ */
+public class DocumentParsedResult {
+
+    private String fileName;
+    private String preferredText;
+    private String markdown;
+    private String plainText;
+    private List<DocumentPage> pages = new ArrayList<DocumentPage>();
+    private List<DocumentBlock> blocks = new ArrayList<DocumentBlock>();
+    private List<DocumentTable> tables = new ArrayList<DocumentTable>();
+    private List<DocumentImage> images = new ArrayList<DocumentImage>();
+    private List<String> warnings = new ArrayList<String>();
+    private Map<String, Object> metadata = new LinkedHashMap<String, Object>();
+    private DocumentParseArtifacts artifacts = new DocumentParseArtifacts();
+
+    public String getFileName() {
+        return fileName;
+    }
+
+    public void setFileName(String fileName) {
+        this.fileName = fileName;
+    }
+
+    public String getPreferredText() {
+        return preferredText;
+    }
+
+    public void setPreferredText(String preferredText) {
+        this.preferredText = preferredText;
+    }
+
+    public String getMarkdown() {
+        return markdown;
+    }
+
+    public void setMarkdown(String markdown) {
+        this.markdown = markdown;
+    }
+
+    public String getPlainText() {
+        return plainText;
+    }
+
+    public void setPlainText(String plainText) {
+        this.plainText = plainText;
+    }
+
+    public List<DocumentPage> getPages() {
+        return pages;
+    }
+
+    public void setPages(List<DocumentPage> pages) {
+        this.pages = pages == null ? new ArrayList<DocumentPage>() : pages;
+    }
+
+    public List<DocumentBlock> getBlocks() {
+        return blocks;
+    }
+
+    public void setBlocks(List<DocumentBlock> blocks) {
+        this.blocks = blocks == null ? new ArrayList<DocumentBlock>() : blocks;
+    }
+
+    public List<DocumentTable> getTables() {
+        return tables;
+    }
+
+    public void setTables(List<DocumentTable> tables) {
+        this.tables = tables == null ? new ArrayList<DocumentTable>() : tables;
+    }
+
+    public List<DocumentImage> getImages() {
+        return images;
+    }
+
+    public void setImages(List<DocumentImage> images) {
+        this.images = images == null ? new ArrayList<DocumentImage>() : images;
+    }
+
+    public List<String> getWarnings() {
+        return warnings;
+    }
+
+    public void setWarnings(List<String> warnings) {
+        this.warnings = warnings == null ? new ArrayList<String>() : warnings;
+    }
+
+    public Map<String, Object> getMetadata() {
+        return metadata;
+    }
+
+    public void setMetadata(Map<String, Object> metadata) {
+        this.metadata = metadata == null ? new LinkedHashMap<String, Object>() : metadata;
+    }
+
+    public DocumentParseArtifacts getArtifacts() {
+        return artifacts;
+    }
+
+    public void setArtifacts(DocumentParseArtifacts artifacts) {
+        this.artifacts = artifacts == null ? new DocumentParseArtifacts() : artifacts;
+    }
+}
--- a/easyflow-modules/easyflow-module-ai/src/main/java/tech/easyflow/ai/document/model/DocumentSourceRef.java
+++ b/easyflow-modules/easyflow-module-ai/src/main/java/tech/easyflow/ai/document/model/DocumentSourceRef.java
@@ -0,0 +1,106 @@
+package tech.easyflow.ai.document.model;
+
+/**
+ * 统一文档源引用。
+ *
+ * <p>该模型用于屏蔽业务模块和底层解析框架的文件输入差异，业务方只需要描述文件来自哪里，
+ * 具体由桥接层负责加载字节内容并转成统一解析请求。</p>
+ *
+ * @author Codex
+ * @since 2026-04-14
+ */
+public class DocumentSourceRef {
+
+    private String fileName;
+    private String filePath;
+    private String contentType;
+    private Long size;
+    private String url;
+    private byte[] contentBytes;
+
+    /**
+     * 创建基于文件存储路径的文档源。
+     *
+     * @param filePath 存储路径
+     * @return 文档源
+     */
+    public static DocumentSourceRef ofPath(String filePath) {
+        DocumentSourceRef sourceRef = new DocumentSourceRef();
+        sourceRef.setFilePath(filePath);
+        return sourceRef;
+    }
+
+    /**
+     * 创建基于内存字节的文档源。
+     *
+     * @param fileName 文件名
+     * @param contentBytes 文件字节
+     * @return 文档源
+     */
+    public static DocumentSourceRef ofBytes(String fileName, byte[] contentBytes) {
+        DocumentSourceRef sourceRef = new DocumentSourceRef();
+        sourceRef.setFileName(fileName);
+        sourceRef.setContentBytes(contentBytes);
+        return sourceRef;
+    }
+
+    /**
+     * 创建基于 URL 的文档源。
+     *
+     * @param url 文件 URL
+     * @return 文档源
+     */
+    public static DocumentSourceRef ofUrl(String url) {
+        DocumentSourceRef sourceRef = new DocumentSourceRef();
+        sourceRef.setUrl(url);
+        return sourceRef;
+    }
+
+    public String getFileName() {
+        return fileName;
+    }
+
+    public void setFileName(String fileName) {
+        this.fileName = fileName;
+    }
+
+    public String getFilePath() {
+        return filePath;
+    }
+
+    public void setFilePath(String filePath) {
+        this.filePath = filePath;
+    }
+
+    public String getContentType() {
+        return contentType;
+    }
+
+    public void setContentType(String contentType) {
+        this.contentType = contentType;
+    }
+
+    public Long getSize() {
+        return size;
+    }
+
+    public void setSize(Long size) {
+        this.size = size;
+    }
+
+    public String getUrl() {
+        return url;
+    }
+
+    public void setUrl(String url) {
+        this.url = url;
+    }
+
+    public byte[] getContentBytes() {
+        return contentBytes;
+    }
+
+    public void setContentBytes(byte[] contentBytes) {
+        this.contentBytes = contentBytes;
+    }
+}
--- a/easyflow-modules/easyflow-module-ai/src/main/java/tech/easyflow/ai/document/service/DocumentParseBridgeService.java
+++ b/easyflow-modules/easyflow-module-ai/src/main/java/tech/easyflow/ai/document/service/DocumentParseBridgeService.java
@@ -0,0 +1,67 @@
+package tech.easyflow.ai.document.service;
+
+import tech.easyflow.ai.document.model.DocumentParseScenario;
+import tech.easyflow.ai.document.model.DocumentSourceRef;
+import tech.easyflow.ai.document.model.DocumentParseTaskInfo;
+import tech.easyflow.ai.document.model.DocumentParseTaskStatus;
+import tech.easyflow.ai.document.model.DocumentParsedResult;
+
+/**
+ * 统一文档解析桥接门面。
+ *
+ * <p>业务模块通过该门面使用文档解析能力，而不是直接依赖 easy-agents 的原始请求和结果模型。</p>
+ *
+ * @author Codex
+ * @since 2026-04-14
+ */
+public interface DocumentParseBridgeService {
+
+    /**
+     * 同步解析单文档。
+     *
+     * @param source 文档源
+     * @param scenario 解析场景
+     * @return 标准化解析结果
+     */
+    DocumentParsedResult parse(DocumentSourceRef source, DocumentParseScenario scenario);
+
+    /**
+     * 异步提交单文档解析任务。
+     *
+     * @param source 文档源
+     * @param scenario 解析场景
+     * @return 异步任务状态
+     */
+    DocumentParseTaskStatus submit(DocumentSourceRef source, DocumentParseScenario scenario);
+
+    /**
+     * 查询异步任务状态。
+     *
+     * @param taskId 任务 ID
+     * @return 异步任务状态
+     */
+    DocumentParseTaskStatus queryTask(String taskId);
+
+    /**
+     * 获取异步任务最终结果。
+     *
+     * <p>该方法面向“结果读取”语义，底层 provider 可能在内部等待任务完成后再返回最终结果，
+     * 因此不适合直接作为轻量状态轮询接口；如果业务需要统一查看“当前状态 + 已完成结果”，
+     * 应优先使用 {@link #queryTaskInfo(String)}。</p>
+     *
+     * @param taskId 任务 ID
+     * @return 标准化解析结果
+     */
+    DocumentParsedResult queryResult(String taskId);
+
+    /**
+     * 聚合查询异步任务信息。
+     *
+     * <p>当任务仍在处理中时仅返回状态；当任务已完成时会附带标准化结果。
+     * 该方法适合用于页面或业务侧统一读取“当前状态 + 可用结果”。</p>
+     *
+     * @param taskId 任务 ID
+     * @return 聚合任务信息
+     */
+    DocumentParseTaskInfo queryTaskInfo(String taskId);
+}
--- a/easyflow-modules/easyflow-module-ai/src/main/java/tech/easyflow/ai/document/service/impl/DocumentParseBridgeServiceImpl.java
+++ b/easyflow-modules/easyflow-module-ai/src/main/java/tech/easyflow/ai/document/service/impl/DocumentParseBridgeServiceImpl.java
@@ -0,0 +1,170 @@
+package tech.easyflow.ai.document.service.impl;
+
+import com.easyagents.document.core.DocumentParseService;
+import com.easyagents.document.core.model.ParseResponse;
+import com.easyagents.document.core.model.ParseResult;
+import com.easyagents.document.core.model.ParseTaskInfo;
+import com.easyagents.document.core.model.ParseTaskStatus;
+import org.springframework.lang.Nullable;
+import org.springframework.stereotype.Service;
+import org.springframework.util.StringUtils;
+import tech.easyflow.ai.document.exception.DocumentParseBridgeException;
+import tech.easyflow.ai.document.model.DocumentParseScenario;
+import tech.easyflow.ai.document.model.DocumentSourceRef;
+import tech.easyflow.ai.document.model.DocumentParseTaskInfo;
+import tech.easyflow.ai.document.model.DocumentParseTaskStatus;
+import tech.easyflow.ai.document.model.DocumentParsedResult;
+import tech.easyflow.ai.document.service.DocumentParseBridgeService;
+import tech.easyflow.ai.document.support.DocumentSourceLoader;
+import tech.easyflow.ai.document.support.DocumentParseRequestFactory;
+import tech.easyflow.ai.document.support.DocumentParseResultMapper;
+import tech.easyflow.ai.document.support.LoadedDocumentSource;
+import tech.easyflow.ai.utils.DocUtil;
+
+/**
+ * 统一文档解析桥接门面默认实现。
+ *
+ * @author Codex
+ * @since 2026-04-14
+ */
+@Service
+public class DocumentParseBridgeServiceImpl implements DocumentParseBridgeService {
+
+    @Nullable
+    private final DocumentParseService documentParseService;
+    private final DocumentSourceLoader documentSourceLoader;
+    private final DocumentParseRequestFactory parseRequestFactory;
+    private final DocumentParseResultMapper parseResultMapper;
+
+    public DocumentParseBridgeServiceImpl(@Nullable DocumentParseService documentParseService,
+                                          DocumentSourceLoader documentSourceLoader,
+                                          DocumentParseRequestFactory parseRequestFactory,
+                                          DocumentParseResultMapper parseResultMapper) {
+        this.documentParseService = documentParseService;
+        this.documentSourceLoader = documentSourceLoader;
+        this.parseRequestFactory = parseRequestFactory;
+        this.parseResultMapper = parseResultMapper;
+    }
+
+    /**
+     * {@inheritDoc}
+     */
+    @Override
+    public DocumentParsedResult parse(DocumentSourceRef source, DocumentParseScenario scenario) {
+        try {
+            LoadedDocumentSource loadedSource = preparePdfSource(source);
+            ParseResponse response = requireService().parse(parseRequestFactory.build(loadedSource, scenario));
+            return parseResultMapper.map(extractSingleResult(response, false));
+        } catch (DocumentParseBridgeException e) {
+            throw e;
+        } catch (Exception e) {
+            throw DocumentParseBridgeException.parseFailed("同步文档解析失败", e);
+        }
+    }
+
+    /**
+     * {@inheritDoc}
+     */
+    @Override
+    public DocumentParseTaskStatus submit(DocumentSourceRef source, DocumentParseScenario scenario) {
+        try {
+            LoadedDocumentSource loadedSource = preparePdfSource(source);
+            ParseTaskStatus taskStatus = requireService().submit(parseRequestFactory.build(loadedSource, scenario));
+            return parseResultMapper.map(taskStatus);
+        } catch (DocumentParseBridgeException e) {
+            throw e;
+        } catch (Exception e) {
+            throw DocumentParseBridgeException.taskFailed("提交异步文档解析任务失败", e);
+        }
+    }
+
+    /**
+     * {@inheritDoc}
+     */
+    @Override
+    public DocumentParseTaskStatus queryTask(String taskId) {
+        if (!StringUtils.hasText(taskId)) {
+            throw DocumentParseBridgeException.taskFailed("taskId 不能为空");
+        }
+        try {
+            return parseResultMapper.map(requireService().queryTask(taskId));
+        } catch (DocumentParseBridgeException e) {
+            throw e;
+        } catch (Exception e) {
+            throw DocumentParseBridgeException.taskFailed("查询异步文档解析任务状态失败", e);
+        }
+    }
+
+    /**
+     * {@inheritDoc}
+     */
+    @Override
+    public DocumentParsedResult queryResult(String taskId) {
+        if (!StringUtils.hasText(taskId)) {
+            throw DocumentParseBridgeException.resultFetchFailed("taskId 不能为空");
+        }
+        try {
+            ParseResponse response = requireService().queryResult(taskId);
+            return parseResultMapper.map(extractSingleResult(response, true));
+        } catch (DocumentParseBridgeException e) {
+            throw e;
+        } catch (Exception e) {
+            throw DocumentParseBridgeException.resultFetchFailed("获取异步文档解析结果失败", e);
+        }
+    }
+
+    /**
+     * {@inheritDoc}
+     */
+    @Override
+    public DocumentParseTaskInfo queryTaskInfo(String taskId) {
+        if (!StringUtils.hasText(taskId)) {
+            throw DocumentParseBridgeException.taskFailed("taskId 不能为空");
+        }
+        try {
+            ParseTaskInfo taskInfo = requireService().queryTaskInfo(taskId);
+            return parseResultMapper.map(taskInfo);
+        } catch (DocumentParseBridgeException e) {
+            throw e;
+        } catch (Exception e) {
+            throw DocumentParseBridgeException.taskFailed("聚合查询异步文档解析任务信息失败", e);
+        }
+    }
+
+    private DocumentParseService requireService() {
+        if (documentParseService == null) {
+            throw DocumentParseBridgeException.serviceNotEnabled();
+        }
+        return documentParseService;
+    }
+
+    private LoadedDocumentSource preparePdfSource(DocumentSourceRef source) {
+        LoadedDocumentSource loadedSource = documentSourceLoader.load(source);
+        if (!isPdf(loadedSource)) {
+            throw DocumentParseBridgeException.unsupportedSource("统一文档解析桥接首版仅支持 PDF 文件");
+        }
+        return loadedSource;
+    }
+
+    private boolean isPdf(LoadedDocumentSource loadedSource) {
+        String contentType = loadedSource.getContentType();
+        if (StringUtils.hasText(contentType) && contentType.toLowerCase().contains("pdf")) {
+            return true;
+        }
+        String fileName = loadedSource.getFileName();
+        if (!StringUtils.hasText(fileName) || !fileName.contains(".")) {
+            return false;
+        }
+        return "pdf".equals(DocUtil.normalizeSuffix(DocUtil.getSuffix(fileName)));
+    }
+
+    private ParseResult extractSingleResult(ParseResponse response, boolean resultFetchPhase) {
+        if (response == null || response.getResults() == null || response.getResults().isEmpty()) {
+            if (resultFetchPhase) {
+                throw DocumentParseBridgeException.resultFetchFailed("异步文档解析结果为空");
+            }
+            throw DocumentParseBridgeException.parseFailed("同步文档解析结果为空");
+        }
+        return response.getResults().get(0);
+    }
+}
--- a/easyflow-modules/easyflow-module-ai/src/main/java/tech/easyflow/ai/document/support/DocumentParseRequestFactory.java
+++ b/easyflow-modules/easyflow-module-ai/src/main/java/tech/easyflow/ai/document/support/DocumentParseRequestFactory.java
@@ -0,0 +1,71 @@
+package tech.easyflow.ai.document.support;
+
+import com.easyagents.document.core.model.ParseFile;
+import com.easyagents.document.core.model.ParseRequest;
+import org.springframework.stereotype.Component;
+import tech.easyflow.ai.document.exception.DocumentParseBridgeException;
+import tech.easyflow.ai.document.model.DocumentParseScenario;
+
+/**
+ * easy-agents 解析请求工厂。
+ *
+ * <p>负责把 easyflow 业务场景预设映射为底层统一解析请求。</p>
+ *
+ * @author Codex
+ * @since 2026-04-14
+ */
+@Component
+public class DocumentParseRequestFactory {
+
+    /**
+     * 构建统一解析请求。
+     *
+     * @param source 已加载文档源
+     * @param scenario 解析场景
+     * @return 统一解析请求
+     */
+    public ParseRequest build(LoadedDocumentSource source, DocumentParseScenario scenario) {
+        if (source == null || source.getContentBytes() == null || source.getContentBytes().length == 0) {
+            throw DocumentParseBridgeException.requestBuildFailed("文档源内容为空，无法构建解析请求");
+        }
+        if (scenario == null) {
+            throw DocumentParseBridgeException.requestBuildFailed("解析场景不能为空");
+        }
+        ParseRequest request = new ParseRequest();
+        // 保持为空，交由 easy-agents provider 按环境配置回填默认值。
+        request.setParseMethod(null);
+        request.setFormulaEnabled(null);
+        request.setTableEnabled(null);
+        request.addFile(ParseFile.of(source.getFileName(), source.getContentBytes(), source.getContentType()));
+        applyScenario(request, scenario);
+        return request;
+    }
+
+    private void applyScenario(ParseRequest request, DocumentParseScenario scenario) {
+        switch (scenario) {
+            case WORKFLOW_TEXT:
+                request.setReturnMarkdown(Boolean.TRUE);
+                request.setReturnMiddleJson(Boolean.FALSE);
+                request.setReturnContentList(Boolean.FALSE);
+                request.setReturnModelOutput(Boolean.FALSE);
+                request.setReturnImages(Boolean.FALSE);
+                break;
+            case KNOWLEDGE_IMPORT:
+                request.setReturnMarkdown(Boolean.TRUE);
+                request.setReturnMiddleJson(Boolean.TRUE);
+                request.setReturnContentList(Boolean.TRUE);
+                request.setReturnModelOutput(Boolean.FALSE);
+                request.setReturnImages(Boolean.TRUE);
+                break;
+            case FULL_ARTIFACTS:
+                request.setReturnMarkdown(Boolean.TRUE);
+                request.setReturnMiddleJson(Boolean.TRUE);
+                request.setReturnContentList(Boolean.TRUE);
+                request.setReturnModelOutput(Boolean.TRUE);
+                request.setReturnImages(Boolean.TRUE);
+                break;
+            default:
+                throw DocumentParseBridgeException.requestBuildFailed("不支持的文档解析场景: " + scenario);
+        }
+    }
+}
--- a/easyflow-modules/easyflow-module-ai/src/main/java/tech/easyflow/ai/document/support/DocumentParseResultMapper.java
+++ b/easyflow-modules/easyflow-module-ai/src/main/java/tech/easyflow/ai/document/support/DocumentParseResultMapper.java
@@ -0,0 +1,128 @@
+package tech.easyflow.ai.document.support;
+
+import com.easyagents.document.core.model.ParseArtifacts;
+import com.easyagents.document.core.model.ParseResult;
+import com.easyagents.document.core.model.ParseTaskInfo;
+import com.easyagents.document.core.model.ParseTaskStatus;
+import org.springframework.stereotype.Component;
+import org.springframework.util.StringUtils;
+import tech.easyflow.ai.document.model.DocumentParseArtifacts;
+import tech.easyflow.ai.document.model.DocumentParseTaskInfo;
+import tech.easyflow.ai.document.model.DocumentParseTaskStatus;
+import tech.easyflow.ai.document.model.DocumentParsedResult;
+
+/**
+ * easy-agents 结果映射器。
+ *
+ * <p>负责把底层解析结果转换为 easyflow 侧稳定 DTO，并统一 preferredText 规则。</p>
+ *
+ * @author Codex
+ * @since 2026-04-14
+ */
+@Component
+public class DocumentParseResultMapper {
+
+    /**
+     * 映射单文件解析结果。
+     *
+     * @param parseResult 底层结果
+     * @return easyflow 结果
+     */
+    public DocumentParsedResult map(ParseResult parseResult) {
+        DocumentParsedResult document = new DocumentParsedResult();
+        if (parseResult == null) {
+            return document;
+        }
+        document.setFileName(parseResult.getFileName());
+        document.setMarkdown(parseResult.getMarkdown());
+        document.setPlainText(parseResult.getPlainText());
+        document.setPreferredText(resolvePreferredText(parseResult));
+        document.setPages(parseResult.getPages());
+        document.setBlocks(parseResult.getBlocks());
+        document.setTables(parseResult.getTables());
+        document.setImages(parseResult.getImages());
+        document.setWarnings(parseResult.getWarnings());
+        document.setMetadata(parseResult.getMetadata());
+        document.setArtifacts(mapArtifacts(parseResult.getArtifacts()));
+        return document;
+    }
+
+    /**
+     * 映射异步任务状态。
+     *
+     * @param taskStatus 底层任务状态
+     * @return easyflow 任务状态
+     */
+    public DocumentParseTaskStatus map(ParseTaskStatus taskStatus) {
+        DocumentParseTaskStatus status = new DocumentParseTaskStatus();
+        if (taskStatus == null) {
+            return status;
+        }
+        status.setTaskId(taskStatus.getTaskId());
+        status.setStatus(taskStatus.getStatus());
+        status.setBackend(taskStatus.getBackend());
+        status.setFileNames(taskStatus.getFileNames());
+        status.setCreatedAt(taskStatus.getCreatedAt());
+        status.setStartedAt(taskStatus.getStartedAt());
+        status.setCompletedAt(taskStatus.getCompletedAt());
+        status.setError(taskStatus.getError());
+        status.setStatusUrl(taskStatus.getStatusUrl());
+        status.setResultUrl(taskStatus.getResultUrl());
+        status.setQueuedAhead(taskStatus.getQueuedAhead());
+        return status;
+    }
+
+    /**
+     * 映射任务聚合查询结果。
+     *
+     * @param taskInfo 底层任务聚合结果
+     * @return easyflow 聚合任务结果
+     */
+    public DocumentParseTaskInfo map(ParseTaskInfo taskInfo) {
+        DocumentParseTaskInfo mapped = new DocumentParseTaskInfo();
+        if (taskInfo == null) {
+            return mapped;
+        }
+        fillTaskStatus(mapped, taskInfo);
+        if (taskInfo.getResult() != null
+            && taskInfo.getResult().getResults() != null
+            && !taskInfo.getResult().getResults().isEmpty()) {
+            mapped.setResult(map(taskInfo.getResult().getResults().get(0)));
+        }
+        return mapped;
+    }
+
+    private void fillTaskStatus(DocumentParseTaskStatus status, ParseTaskStatus taskStatus) {
+        status.setTaskId(taskStatus.getTaskId());
+        status.setStatus(taskStatus.getStatus());
+        status.setBackend(taskStatus.getBackend());
+        status.setFileNames(taskStatus.getFileNames());
+        status.setCreatedAt(taskStatus.getCreatedAt());
+        status.setStartedAt(taskStatus.getStartedAt());
+        status.setCompletedAt(taskStatus.getCompletedAt());
+        status.setError(taskStatus.getError());
+        status.setStatusUrl(taskStatus.getStatusUrl());
+        status.setResultUrl(taskStatus.getResultUrl());
+        status.setQueuedAhead(taskStatus.getQueuedAhead());
+    }
+
+    private String resolvePreferredText(ParseResult parseResult) {
+        if (StringUtils.hasText(parseResult.getMarkdown())) {
+            return parseResult.getMarkdown();
+        }
+        return parseResult.getPlainText();
+    }
+
+    private DocumentParseArtifacts mapArtifacts(ParseArtifacts artifacts) {
+        DocumentParseArtifacts mappedArtifacts = new DocumentParseArtifacts();
+        if (artifacts == null) {
+            return mappedArtifacts;
+        }
+        mappedArtifacts.setMiddleJson(artifacts.getMiddleJson());
+        mappedArtifacts.setContentList(artifacts.getContentList());
+        mappedArtifacts.setModelOutput(artifacts.getModelOutput());
+        mappedArtifacts.setExtraJsonArtifacts(artifacts.getExtraJsonArtifacts());
+        mappedArtifacts.setExtraBinaryArtifacts(artifacts.getExtraBinaryArtifacts());
+        return mappedArtifacts;
+    }
+}
--- a/easyflow-modules/easyflow-module-ai/src/main/java/tech/easyflow/ai/document/support/DocumentSourceLoader.java
+++ b/easyflow-modules/easyflow-module-ai/src/main/java/tech/easyflow/ai/document/support/DocumentSourceLoader.java
@@ -0,0 +1,146 @@
+package tech.easyflow.ai.document.support;
+
+import cn.hutool.http.HttpUtil;
+import org.springframework.beans.factory.annotation.Qualifier;
+import org.springframework.stereotype.Component;
+import org.springframework.util.StringUtils;
+import tech.easyflow.ai.document.exception.DocumentParseBridgeException;
+import tech.easyflow.ai.document.model.DocumentSourceRef;
+import tech.easyflow.common.filestorage.FileStorageService;
+import tech.easyflow.common.filestorage.utils.PathGeneratorUtil;
+
+import java.io.IOException;
+import java.io.InputStream;
+import java.net.URLConnection;
+
+/**
+ * 文档源加载器。
+ *
+ * <p>负责把不同来源的文件引用统一转换为内存字节和标准文件元数据。</p>
+ *
+ * @author Codex
+ * @since 2026-04-14
+ */
+@Component
+public class DocumentSourceLoader {
+
+    private final FileStorageService fileStorageService;
+
+    public DocumentSourceLoader(@Qualifier("default") FileStorageService fileStorageService) {
+        this.fileStorageService = fileStorageService;
+    }
+
+    /**
+     * 加载文档源。
+     *
+     * @param sourceRef easyflow 文档源
+     * @return 内部已加载文档对象
+     */
+    public LoadedDocumentSource load(DocumentSourceRef sourceRef) {
+        if (sourceRef == null) {
+            throw DocumentParseBridgeException.unsupportedSource("文档源不能为空");
+        }
+        if (hasContentBytes(sourceRef)) {
+            return buildLoadedSource(
+                resolveFileName(sourceRef),
+                resolveContentType(sourceRef, resolveFileName(sourceRef)),
+                resolveSize(sourceRef, sourceRef.getContentBytes().length),
+                sourceRef.getContentBytes()
+            );
+        }
+        if (StringUtils.hasText(sourceRef.getFilePath())) {
+            if (isRemoteUrl(sourceRef.getFilePath())) {
+                return loadFromRemoteValue(sourceRef, sourceRef.getFilePath());
+            }
+            return loadFromFilePath(sourceRef);
+        }
+        if (StringUtils.hasText(sourceRef.getUrl())) {
+            return loadFromUrl(sourceRef);
+        }
+        throw DocumentParseBridgeException.unsupportedSource("文档源缺少 filePath、url 或 contentBytes");
+    }
+
+    private LoadedDocumentSource loadFromFilePath(DocumentSourceRef sourceRef) {
+        String fileName = resolveFileName(sourceRef);
+        try (InputStream inputStream = fileStorageService.readStream(sourceRef.getFilePath())) {
+            byte[] contentBytes = inputStream.readAllBytes();
+            long actualSize = sourceRef.getSize() != null ? sourceRef.getSize() : fileStorageService.getFileSize(sourceRef.getFilePath());
+            return buildLoadedSource(
+                fileName,
+                resolveContentType(sourceRef, fileName),
+                resolveSize(sourceRef, actualSize),
+                contentBytes
+            );
+        } catch (IOException e) {
+            throw DocumentParseBridgeException.sourceLoadFailed(
+                "读取文档存储文件失败: " + sourceRef.getFilePath(),
+                e
+            );
+        }
+    }
+
+    private LoadedDocumentSource loadFromUrl(DocumentSourceRef sourceRef) {
+        return loadFromRemoteValue(sourceRef, sourceRef.getUrl());
+    }
+
+    private LoadedDocumentSource loadFromRemoteValue(DocumentSourceRef sourceRef, String remoteUrl) {
+        String fileName = resolveFileName(sourceRef);
+        try {
+            byte[] contentBytes = HttpUtil.downloadBytes(remoteUrl);
+            return buildLoadedSource(
+                fileName,
+                resolveContentType(sourceRef, fileName),
+                resolveSize(sourceRef, contentBytes.length),
+                contentBytes
+            );
+        } catch (Exception e) {
+            throw DocumentParseBridgeException.sourceLoadFailed(
+                "下载文档 URL 失败: " + remoteUrl,
+                e
+            );
+        }
+    }
+
+    private LoadedDocumentSource buildLoadedSource(String fileName, String contentType, Long size, byte[] contentBytes) {
+        LoadedDocumentSource loadedSource = new LoadedDocumentSource();
+        loadedSource.setFileName(fileName);
+        loadedSource.setContentType(contentType);
+        loadedSource.setSize(size);
+        loadedSource.setContentBytes(contentBytes);
+        return loadedSource;
+    }
+
+    private String resolveFileName(DocumentSourceRef sourceRef) {
+        if (StringUtils.hasText(sourceRef.getFileName())) {
+            return PathGeneratorUtil.getPureFileName(sourceRef.getFileName());
+        }
+        if (StringUtils.hasText(sourceRef.getFilePath())) {
+            return PathGeneratorUtil.getPureFileName(sourceRef.getFilePath());
+        }
+        if (StringUtils.hasText(sourceRef.getUrl())) {
+            String pureName = PathGeneratorUtil.getPureFileName(sourceRef.getUrl());
+            int queryIndex = pureName.indexOf('?');
+            return queryIndex >= 0 ? pureName.substring(0, queryIndex) : pureName;
+        }
+        throw DocumentParseBridgeException.unsupportedSource("文档源缺少可用文件名");
+    }
+
+    private String resolveContentType(DocumentSourceRef sourceRef, String fileName) {
+        if (StringUtils.hasText(sourceRef.getContentType())) {
+            return sourceRef.getContentType();
+        }
+        return URLConnection.guessContentTypeFromName(fileName);
+    }
+
+    private Long resolveSize(DocumentSourceRef sourceRef, long fallbackSize) {
+        return sourceRef.getSize() != null ? sourceRef.getSize() : fallbackSize;
+    }
+
+    private boolean hasContentBytes(DocumentSourceRef sourceRef) {
+        return sourceRef.getContentBytes() != null && sourceRef.getContentBytes().length > 0;
+    }
+
+    private boolean isRemoteUrl(String value) {
+        return value.startsWith("http://") || value.startsWith("https://");
+    }
+}
--- a/easyflow-modules/easyflow-module-ai/src/main/java/tech/easyflow/ai/document/support/LoadedDocumentSource.java
+++ b/easyflow-modules/easyflow-module-ai/src/main/java/tech/easyflow/ai/document/support/LoadedDocumentSource.java
@@ -0,0 +1,49 @@
+package tech.easyflow.ai.document.support;
+
+/**
+ * 桥接层内部已加载文档源。
+ *
+ * <p>该对象只在桥接层内部流转，用于承接已解析出的文件名、类型和字节内容。</p>
+ *
+ * @author Codex
+ * @since 2026-04-14
+ */
+public class LoadedDocumentSource {
+
+    private String fileName;
+    private String contentType;
+    private Long size;
+    private byte[] contentBytes;
+
+    public String getFileName() {
+        return fileName;
+    }
+
+    public void setFileName(String fileName) {
+        this.fileName = fileName;
+    }
+
+    public String getContentType() {
+        return contentType;
+    }
+
+    public void setContentType(String contentType) {
+        this.contentType = contentType;
+    }
+
+    public Long getSize() {
+        return size;
+    }
+
+    public void setSize(Long size) {
+        this.size = size;
+    }
+
+    public byte[] getContentBytes() {
+        return contentBytes;
+    }
+
+    public void setContentBytes(byte[] contentBytes) {
+        this.contentBytes = contentBytes;
+    }
+}
--- a/easyflow-modules/easyflow-module-ai/src/main/java/tech/easyflow/ai/node/DocNode.java
+++ b/easyflow-modules/easyflow-module-ai/src/main/java/tech/easyflow/ai/node/DocNode.java
@@ -4,28 +4,35 @@ import com.easyagents.core.util.StringUtil;
 import com.easyagents.flow.core.chain.Chain;
 import com.easyagents.flow.core.chain.Parameter;
 import com.easyagents.flow.core.node.BaseNode;
-import org.slf4j.Logger;
-import org.slf4j.LoggerFactory;
-import tech.easyflow.ai.utils.DocUtil;
 import tech.easyflow.common.util.SpringContextUtil;

-import java.io.ByteArrayInputStream;
 import java.util.HashMap;
 import java.util.List;
 import java.util.Map;

+/**
+ * 工作流文件内容提取节点。
+ *
+ * <p>节点输入为统一文件对象，PDF 交给统一文档解析桥接服务，
+ * 其他类型继续走默认文档读取器。</p>
+ *
+ * @author Codex
+ * @since 2026-04-14
+ */
 public class DocNode extends BaseNode {

-    private static final Logger log = LoggerFactory.getLogger(DocNode.class);
-
+    /**
+     * 执行文件内容提取。
+     *
+     * @param chain 当前流程链
+     * @return 节点输出
+     */
    @Override
    public Map<String, Object> execute(Chain chain) {
        Map<String, Object> map = chain.getState().resolveParameters(this);
        Map<String, Object> res = new HashMap<>();
-        String url = map.get("fileUrl").toString();
-        byte[] bytes = DocUtil.downloadFile(url);
-        ReaderManager manager = SpringContextUtil.getBean(ReaderManager.class);
-        String docContent = manager.getReader().read(DocUtil.getFileNameByUrl(url), new ByteArrayInputStream(bytes));
+        DocNodeFileContentExtractor extractor = SpringContextUtil.getBean(DocNodeFileContentExtractor.class);
+        String docContent = extractor.extract(map.get("file"));

        String key = "content";
        List<Parameter> outputDefs = getOutputDefs();
--- a/easyflow-modules/easyflow-module-ai/src/main/java/tech/easyflow/ai/node/DocNodeFileContentExtractor.java
+++ b/easyflow-modules/easyflow-module-ai/src/main/java/tech/easyflow/ai/node/DocNodeFileContentExtractor.java
@@ -0,0 +1,170 @@
+package tech.easyflow.ai.node;
+
+import cn.hutool.http.HttpUtil;
+import org.springframework.beans.factory.annotation.Qualifier;
+import org.springframework.stereotype.Component;
+import tech.easyflow.ai.document.model.DocumentParseScenario;
+import tech.easyflow.ai.document.model.DocumentParsedResult;
+import tech.easyflow.ai.document.model.DocumentSourceRef;
+import tech.easyflow.ai.document.service.DocumentParseBridgeService;
+import tech.easyflow.ai.utils.DocUtil;
+import tech.easyflow.common.filestorage.FileStorageService;
+import tech.easyflow.common.util.StringUtil;
+import tech.easyflow.common.web.exceptions.BusinessException;
+
+import java.io.IOException;
+import java.io.InputStream;
+import java.util.Map;
+
+/**
+ * {@link DocNode} 文件内容提取器。
+ *
+ * <p>负责把工作流运行态中的文件对象转换为统一文档源，并根据文件类型选择
+ * 统一文档解析桥接服务或默认读取器。</p>
+ *
+ * @author Codex
+ * @since 2026-04-14
+ */
+@Component
+public class DocNodeFileContentExtractor {
+
+    private final DocumentParseBridgeService documentParseBridgeService;
+    private final FileStorageService fileStorageService;
+    private final ReaderManager readerManager;
+
+    /**
+     * 创建文件内容提取器。
+     *
+     * @param documentParseBridgeService 统一文档解析桥接服务
+     * @param fileStorageService 文件存储服务
+     * @param readerManager 默认读取器管理器
+     */
+    public DocNodeFileContentExtractor(DocumentParseBridgeService documentParseBridgeService,
+                                       @Qualifier("default") FileStorageService fileStorageService,
+                                       ReaderManager readerManager) {
+        this.documentParseBridgeService = documentParseBridgeService;
+        this.fileStorageService = fileStorageService;
+        this.readerManager = readerManager;
+    }
+
+    /**
+     * 提取文件文本内容。
+     *
+     * @param fileValue 工作流运行态中的文件对象
+     * @return 可供下游节点直接消费的文本
+     */
+    public String extract(Object fileValue) {
+        DocumentSourceRef sourceRef = toDocumentSourceRef(fileValue);
+        validateSourceRef(sourceRef);
+        if (isPdf(sourceRef)) {
+            return extractPdfContent(sourceRef);
+        }
+        return extractDefaultContent(sourceRef);
+    }
+
+    /**
+     * 将运行时文件值转换为统一文档源。
+     *
+     * @param fileValue 运行时文件值
+     * @return 文档源
+     */
+    DocumentSourceRef toDocumentSourceRef(Object fileValue) {
+        if (fileValue instanceof DocumentSourceRef sourceRef) {
+            return sourceRef;
+        }
+        if (!(fileValue instanceof Map<?, ?> fileMap)) {
+            throw new BusinessException("文件输入格式不正确，必须为文件对象");
+        }
+        DocumentSourceRef sourceRef = new DocumentSourceRef();
+        sourceRef.setFileName(asText(fileMap.get("fileName")));
+        sourceRef.setFilePath(asText(fileMap.get("filePath")));
+        sourceRef.setContentType(asText(fileMap.get("contentType")));
+        sourceRef.setUrl(asText(fileMap.get("url")));
+        sourceRef.setSize(asLong(fileMap.get("size")));
+        return sourceRef;
+    }
+
+    private void validateSourceRef(DocumentSourceRef sourceRef) {
+        if (sourceRef == null) {
+            throw new BusinessException("文件输入不能为空");
+        }
+        if (!StringUtil.hasText(sourceRef.getFileName())) {
+            throw new BusinessException("文件输入缺少 fileName");
+        }
+        if (!StringUtil.hasText(sourceRef.getFilePath())) {
+            throw new BusinessException("文件输入缺少 filePath");
+        }
+    }
+
+    private boolean isPdf(DocumentSourceRef sourceRef) {
+        if (StringUtil.hasText(sourceRef.getContentType())
+            && sourceRef.getContentType().toLowerCase().contains("pdf")) {
+            return true;
+        }
+        String fileName = sourceRef.getFileName();
+        if (!StringUtil.hasText(fileName) || !fileName.contains(".")) {
+            return false;
+        }
+        return "pdf".equals(DocUtil.normalizeSuffix(DocUtil.getSuffix(fileName)));
+    }
+
+    private String extractPdfContent(DocumentSourceRef sourceRef) {
+        DocumentParsedResult parsedResult = documentParseBridgeService.parse(sourceRef, DocumentParseScenario.WORKFLOW_TEXT);
+        String preferredText = parsedResult == null ? null : parsedResult.getPreferredText();
+        if (StringUtil.hasText(preferredText)) {
+            return preferredText;
+        }
+        if (parsedResult != null && StringUtil.hasText(parsedResult.getMarkdown())) {
+            return parsedResult.getMarkdown();
+        }
+        if (parsedResult != null && StringUtil.hasText(parsedResult.getPlainText())) {
+            return parsedResult.getPlainText();
+        }
+        throw new BusinessException("PDF 文档解析结果为空");
+    }
+
+    private String extractDefaultContent(DocumentSourceRef sourceRef) {
+        try (InputStream inputStream = openInputStream(sourceRef)) {
+            return readerManager.getReader().read(sourceRef.getFileName(), inputStream);
+        } catch (IOException e) {
+            throw new RuntimeException("读取文件内容失败: " + sourceRef.getFilePath(), e);
+        }
+    }
+
+    private InputStream openInputStream(DocumentSourceRef sourceRef) throws IOException {
+        String filePath = sourceRef.getFilePath();
+        if (StringUtil.hasText(filePath) && isRemoteUrl(filePath)) {
+            byte[] bytes = HttpUtil.downloadBytes(filePath);
+            return new java.io.ByteArrayInputStream(bytes);
+        }
+        if (StringUtil.hasText(filePath)) {
+            return fileStorageService.readStream(filePath);
+        }
+        if (StringUtil.hasText(sourceRef.getUrl())) {
+            byte[] bytes = HttpUtil.downloadBytes(sourceRef.getUrl());
+            return new java.io.ByteArrayInputStream(bytes);
+        }
+        throw new IOException("文件输入缺少可读取路径");
+    }
+
+    private boolean isRemoteUrl(String value) {
+        return value.startsWith("http://") || value.startsWith("https://");
+    }
+
+    private String asText(Object value) {
+        return value == null ? null : String.valueOf(value).trim();
+    }
+
+    private Long asLong(Object value) {
+        if (value == null) {
+            return null;
+        }
+        if (value instanceof Number number) {
+            return number.longValue();
+        }
+        if (value instanceof String text && StringUtil.hasText(text)) {
+            return Long.parseLong(text.trim());
+        }
+        return null;
+    }
+}
--- a/easyflow-modules/easyflow-module-ai/src/test/java/tech/easyflow/ai/document/service/impl/DocumentParseBridgeServiceImplTest.java
+++ b/easyflow-modules/easyflow-module-ai/src/test/java/tech/easyflow/ai/document/service/impl/DocumentParseBridgeServiceImplTest.java
@@ -0,0 +1,190 @@
+package tech.easyflow.ai.document.service.impl;
+
+import com.easyagents.document.core.DocumentParseService;
+import com.easyagents.document.core.model.ParseRequest;
+import com.easyagents.document.core.model.ParseResponse;
+import com.easyagents.document.core.model.ParseResult;
+import com.easyagents.document.core.model.ParseTaskInfo;
+import com.easyagents.document.core.model.ParseTaskStatus;
+import org.junit.Assert;
+import org.junit.Test;
+import tech.easyflow.ai.document.exception.DocumentParseBridgeException;
+import tech.easyflow.ai.document.model.DocumentParseScenario;
+import tech.easyflow.ai.document.model.DocumentSourceRef;
+import tech.easyflow.ai.document.model.DocumentParseTaskInfo;
+import tech.easyflow.ai.document.model.DocumentParseTaskStatus;
+import tech.easyflow.ai.document.model.DocumentParsedResult;
+import tech.easyflow.ai.document.support.DocumentSourceLoader;
+import tech.easyflow.ai.document.support.DocumentParseRequestFactory;
+import tech.easyflow.ai.document.support.DocumentParseResultMapper;
+import tech.easyflow.common.filestorage.FileStorageService;
+
+import java.io.ByteArrayInputStream;
+import java.io.InputStream;
+import java.nio.charset.StandardCharsets;
+import java.util.Collections;
+
+/**
+ * {@link DocumentParseBridgeServiceImpl} 单元测试。
+ *
+ * @author Codex
+ * @since 2026-04-14
+ */
+public class DocumentParseBridgeServiceImplTest {
+
+    /**
+     * 验证同步解析成功透传并返回标准化结果。
+     */
+    @Test
+    public void shouldParseSuccessfully() {
+        FakeDocumentParseService parseService = new FakeDocumentParseService();
+        DocumentParseBridgeServiceImpl bridgeService = buildBridgeService(parseService);
+
+        DocumentParsedResult document = bridgeService.parse(buildSource(), DocumentParseScenario.WORKFLOW_TEXT);
+
+        Assert.assertEquals("# demo", document.getPreferredText());
+        Assert.assertFalse(parseService.lastParseRequest.getReturnMiddleJson());
+        Assert.assertFalse(parseService.lastParseRequest.getReturnImages());
+    }
+
+    /**
+     * 验证异步提交、状态查询和结果查询链路可用。
+     */
+    @Test
+    public void shouldSupportAsyncFlow() {
+        FakeDocumentParseService parseService = new FakeDocumentParseService();
+        DocumentParseBridgeServiceImpl bridgeService = buildBridgeService(parseService);
+
+        DocumentParseTaskStatus taskStatus = bridgeService.submit(buildSource(), DocumentParseScenario.KNOWLEDGE_IMPORT);
+        DocumentParseTaskStatus queriedStatus = bridgeService.queryTask("task-1");
+        DocumentParsedResult queriedResult = bridgeService.queryResult("task-1");
+
+        Assert.assertEquals("task-1", taskStatus.getTaskId());
+        Assert.assertEquals("running", queriedStatus.getStatus());
+        Assert.assertEquals("# demo", queriedResult.getPreferredText());
+    }
+
+    /**
+     * 验证聚合查询在完成状态下会附带标准化结果。
+     */
+    @Test
+    public void shouldQueryTaskInfoSuccessfully() {
+        FakeDocumentParseService parseService = new FakeDocumentParseService();
+        parseService.taskStatusValue = "completed";
+        DocumentParseBridgeServiceImpl bridgeService = buildBridgeService(parseService);
+
+        DocumentParseTaskInfo taskInfo = bridgeService.queryTaskInfo("task-1");
+
+        Assert.assertEquals("completed", taskInfo.getStatus());
+        Assert.assertNotNull(taskInfo.getResult());
+        Assert.assertEquals("# demo", taskInfo.getResult().getPreferredText());
+    }
+
+    /**
+     * 验证缺少底层服务时抛出稳定错误码。
+     */
+    @Test
+    public void shouldThrowWhenServiceDisabled() {
+        DocumentParseBridgeServiceImpl bridgeService = buildBridgeService(null);
+
+        try {
+            bridgeService.parse(buildSource(), DocumentParseScenario.WORKFLOW_TEXT);
+            Assert.fail("expected DocumentParseBridgeException");
+        } catch (DocumentParseBridgeException e) {
+            Assert.assertEquals("service_not_enabled", e.getCode());
+        }
+    }
+
+    private DocumentParseBridgeServiceImpl buildBridgeService(DocumentParseService parseService) {
+        return new DocumentParseBridgeServiceImpl(
+            parseService,
+            new DocumentSourceLoader(new InMemoryFileStorageService()),
+            new DocumentParseRequestFactory(),
+            new DocumentParseResultMapper()
+        );
+    }
+
+    private DocumentSourceRef buildSource() {
+        DocumentSourceRef sourceRef = DocumentSourceRef.ofBytes("demo.pdf", "pdf-data".getBytes(StandardCharsets.UTF_8));
+        sourceRef.setContentType("application/pdf");
+        sourceRef.setSize(8L);
+        return sourceRef;
+    }
+
+    private static class InMemoryFileStorageService implements FileStorageService {
+
+        @Override
+        public String save(org.springframework.web.multipart.MultipartFile file) {
+            return null;
+        }
+
+        @Override
+        public void delete(String path) {
+        }
+
+        @Override
+        public InputStream readStream(String path) {
+            return new ByteArrayInputStream("pdf-data".getBytes(StandardCharsets.UTF_8));
+        }
+
+        @Override
+        public long getFileSize(String path) {
+            return 8L;
+        }
+    }
+
+    private static class FakeDocumentParseService implements DocumentParseService {
+
+        private ParseRequest lastParseRequest;
+        private String taskStatusValue = "running";
+
+        @Override
+        public ParseResponse parse(ParseRequest request) {
+            this.lastParseRequest = request;
+            return buildResponse();
+        }
+
+        @Override
+        public ParseTaskStatus submit(ParseRequest request) {
+            this.lastParseRequest = request;
+            ParseTaskStatus status = new ParseTaskStatus();
+            status.setTaskId("task-1");
+            status.setStatus("submitted");
+            status.setFileNames(Collections.singletonList("demo.pdf"));
+            return status;
+        }
+
+        @Override
+        public ParseTaskStatus queryTask(String taskId) {
+            ParseTaskStatus status = new ParseTaskStatus();
+            status.setTaskId(taskId);
+            status.setStatus(taskStatusValue);
+            status.setFileNames(Collections.singletonList("demo.pdf"));
+            return status;
+        }
+
+        @Override
+        public ParseResponse queryResult(String taskId) {
+            return buildResponse();
+        }
+
+        @Override
+        public ParseTaskInfo queryTaskInfo(String taskId) {
+            ParseTaskInfo taskInfo = ParseTaskInfo.fromStatus(queryTask(taskId));
+            if ("completed".equals(taskStatusValue)) {
+                taskInfo.setResult(buildResponse());
+            }
+            return taskInfo;
+        }
+
+        private ParseResponse buildResponse() {
+            ParseResult result = new ParseResult();
+            result.setFileName("demo.pdf");
+            result.setMarkdown("# demo");
+            result.setPlainText("demo");
+            ParseResponse response = new ParseResponse();
+            response.setResults(Collections.singletonList(result));
+            return response;
+        }
+    }
+}
--- a/easyflow-modules/easyflow-module-ai/src/test/java/tech/easyflow/ai/document/support/DocumentParseRequestFactoryTest.java
+++ b/easyflow-modules/easyflow-module-ai/src/test/java/tech/easyflow/ai/document/support/DocumentParseRequestFactoryTest.java
@@ -0,0 +1,57 @@
+package tech.easyflow.ai.document.support;
+
+import com.easyagents.document.core.model.ParseRequest;
+import org.junit.Assert;
+import org.junit.Test;
+import tech.easyflow.ai.document.model.DocumentParseScenario;
+
+/**
+ * {@link DocumentParseRequestFactory} 单元测试。
+ *
+ * @author Codex
+ * @since 2026-04-14
+ */
+public class DocumentParseRequestFactoryTest {
+
+    /**
+     * 验证工作流文本场景只请求最小文本结果。
+     */
+    @Test
+    public void shouldBuildWorkflowTextScenarioRequest() {
+        DocumentParseRequestFactory factory = new DocumentParseRequestFactory();
+
+        ParseRequest request = factory.build(buildSource(), DocumentParseScenario.WORKFLOW_TEXT);
+
+        Assert.assertNull(request.getParseMethod());
+        Assert.assertNull(request.getFormulaEnabled());
+        Assert.assertNull(request.getTableEnabled());
+        Assert.assertTrue(request.getReturnMarkdown());
+        Assert.assertFalse(request.getReturnMiddleJson());
+        Assert.assertFalse(request.getReturnContentList());
+        Assert.assertFalse(request.getReturnImages());
+    }
+
+    /**
+     * 验证知识库导入场景保留结构化工件。
+     */
+    @Test
+    public void shouldBuildKnowledgeImportScenarioRequest() {
+        DocumentParseRequestFactory factory = new DocumentParseRequestFactory();
+
+        ParseRequest request = factory.build(buildSource(), DocumentParseScenario.KNOWLEDGE_IMPORT);
+
+        Assert.assertTrue(request.getReturnMarkdown());
+        Assert.assertTrue(request.getReturnMiddleJson());
+        Assert.assertTrue(request.getReturnContentList());
+        Assert.assertTrue(request.getReturnImages());
+    }
+
+    private LoadedDocumentSource buildSource() {
+        LoadedDocumentSource source = new LoadedDocumentSource();
+        source.setFileName("demo.pdf");
+        source.setContentType("application/pdf");
+        source.setContentBytes("pdf-data".getBytes());
+        source.setSize(8L);
+        return source;
+    }
+}
--- a/easyflow-modules/easyflow-module-ai/src/test/java/tech/easyflow/ai/document/support/DocumentParseResultMapperTest.java
+++ b/easyflow-modules/easyflow-module-ai/src/test/java/tech/easyflow/ai/document/support/DocumentParseResultMapperTest.java
@@ -0,0 +1,82 @@
+package tech.easyflow.ai.document.support;
+
+import com.easyagents.document.core.model.ParseArtifacts;
+import com.easyagents.document.core.model.ParseResult;
+import com.easyagents.document.core.model.ParseResponse;
+import com.easyagents.document.core.model.ParseTaskInfo;
+import org.junit.Assert;
+import org.junit.Test;
+import tech.easyflow.ai.document.model.DocumentParseTaskInfo;
+import tech.easyflow.ai.document.model.DocumentParsedResult;
+
+import java.util.Collections;
+
+/**
+ * {@link DocumentParseResultMapper} 单元测试。
+ *
+ * @author Codex
+ * @since 2026-04-14
+ */
+public class DocumentParseResultMapperTest {
+
+    /**
+     * 验证 preferredText 按 markdown 优先、plainText 回退。
+     */
+    @Test
+    public void shouldPreferMarkdown() {
+        DocumentParseResultMapper mapper = new DocumentParseResultMapper();
+        ParseResult result = new ParseResult();
+        result.setFileName("demo.pdf");
+        result.setMarkdown("# title");
+        result.setPlainText("plain");
+
+        DocumentParsedResult mapped = mapper.map(result);
+
+        Assert.assertEquals("# title", mapped.getPreferredText());
+    }
+
+    /**
+     * 验证结构化工件被正确映射。
+     */
+    @Test
+    public void shouldMapArtifacts() {
+        DocumentParseResultMapper mapper = new DocumentParseResultMapper();
+        ParseResult result = new ParseResult();
+        ParseArtifacts artifacts = new ParseArtifacts();
+        artifacts.setMiddleJson(Collections.singletonMap("page", 1));
+        artifacts.setContentList(Collections.singletonList("block"));
+        artifacts.setModelOutput(Collections.singletonMap("raw", "ok"));
+        result.setArtifacts(artifacts);
+
+        DocumentParsedResult mapped = mapper.map(result);
+
+        Assert.assertNotNull(mapped.getArtifacts());
+        Assert.assertEquals(Collections.singletonMap("page", 1), mapped.getArtifacts().getMiddleJson());
+        Assert.assertEquals(Collections.singletonList("block"), mapped.getArtifacts().getContentList());
+        Assert.assertEquals(Collections.singletonMap("raw", "ok"), mapped.getArtifacts().getModelOutput());
+    }
+
+    /**
+     * 验证任务聚合结果被正确映射。
+     */
+    @Test
+    public void shouldMapTaskInfo() {
+        DocumentParseResultMapper mapper = new DocumentParseResultMapper();
+        ParseTaskInfo taskInfo = new ParseTaskInfo();
+        taskInfo.setTaskId("task-1");
+        taskInfo.setStatus("completed");
+
+        ParseResult result = new ParseResult();
+        result.setFileName("demo.pdf");
+        result.setMarkdown("# title");
+        ParseResponse response = new ParseResponse();
+        response.setResults(Collections.singletonList(result));
+        taskInfo.setResult(response);
+
+        DocumentParseTaskInfo mapped = mapper.map(taskInfo);
+
+        Assert.assertEquals("task-1", mapped.getTaskId());
+        Assert.assertNotNull(mapped.getResult());
+        Assert.assertEquals("# title", mapped.getResult().getPreferredText());
+    }
+}
--- a/easyflow-modules/easyflow-module-ai/src/test/java/tech/easyflow/ai/document/support/DocumentSourceLoaderTest.java
+++ b/easyflow-modules/easyflow-module-ai/src/test/java/tech/easyflow/ai/document/support/DocumentSourceLoaderTest.java
@@ -0,0 +1,133 @@
+package tech.easyflow.ai.document.support;
+
+import org.junit.Assert;
+import org.junit.Test;
+import tech.easyflow.ai.document.exception.DocumentParseBridgeException;
+import tech.easyflow.ai.document.model.DocumentSourceRef;
+import tech.easyflow.common.filestorage.FileStorageService;
+
+import java.io.ByteArrayInputStream;
+import java.io.IOException;
+import java.io.InputStream;
+import java.net.InetSocketAddress;
+import java.nio.charset.StandardCharsets;
+import com.sun.net.httpserver.HttpServer;
+
+/**
+ * {@link DocumentSourceLoader} 单元测试。
+ *
+ * @author Codex
+ * @since 2026-04-14
+ */
+public class DocumentSourceLoaderTest {
+
+    /**
+     * 验证可从 filePath 正常读取文件内容。
+     *
+     * @throws IOException 测试读取异常
+     */
+    @Test
+    public void shouldLoadContentFromFilePath() throws IOException {
+        DocumentSourceLoader loader = new DocumentSourceLoader(new FakeFileStorageService("demo-pdf".getBytes(StandardCharsets.UTF_8)));
+        DocumentSourceRef sourceRef = DocumentSourceRef.ofPath("/attachment/test/demo.pdf");
+
+        LoadedDocumentSource loadedSource = loader.load(sourceRef);
+
+        Assert.assertEquals("demo.pdf", loadedSource.getFileName());
+        Assert.assertEquals(8L, loadedSource.getSize().longValue());
+        Assert.assertArrayEquals("demo-pdf".getBytes(StandardCharsets.UTF_8), loadedSource.getContentBytes());
+    }
+
+    /**
+     * 验证缺少有效来源时抛出明确异常。
+     */
+    @Test
+    public void shouldThrowWhenSourceMissing() {
+        DocumentSourceLoader loader = new DocumentSourceLoader(new FakeFileStorageService(new byte[0]));
+
+        try {
+            loader.load(new DocumentSourceRef());
+            Assert.fail("expected DocumentParseBridgeException");
+        } catch (DocumentParseBridgeException e) {
+            Assert.assertEquals("unsupported_source", e.getCode());
+        }
+    }
+
+    /**
+     * 验证 filePath 为远端 URL 时不会误走存储读取。
+     */
+    @Test
+    public void shouldPreferRemoteDownloadWhenFilePathIsRemoteUrl() throws IOException {
+        DocumentSourceLoader loader = new DocumentSourceLoader(new FailingFileStorageService());
+        HttpServer server = HttpServer.create(new InetSocketAddress(0), 0);
+        byte[] body = "demo-pdf".getBytes(StandardCharsets.UTF_8);
+        server.createContext("/demo.pdf", exchange -> {
+            exchange.sendResponseHeaders(200, body.length);
+            exchange.getResponseBody().write(body);
+            exchange.close();
+        });
+        server.start();
+        try {
+            DocumentSourceRef sourceRef = new DocumentSourceRef();
+            sourceRef.setFileName("demo.pdf");
+            sourceRef.setFilePath("http://127.0.0.1:" + server.getAddress().getPort() + "/demo.pdf");
+
+            LoadedDocumentSource loadedSource = loader.load(sourceRef);
+
+            Assert.assertEquals("demo.pdf", loadedSource.getFileName());
+            Assert.assertArrayEquals(body, loadedSource.getContentBytes());
+        } finally {
+            server.stop(0);
+        }
+    }
+
+    private static class FakeFileStorageService implements FileStorageService {
+
+        private final byte[] content;
+
+        private FakeFileStorageService(byte[] content) {
+            this.content = content;
+        }
+
+        @Override
+        public String save(org.springframework.web.multipart.MultipartFile file) {
+            return null;
+        }
+
+        @Override
+        public void delete(String path) {
+        }
+
+        @Override
+        public InputStream readStream(String path) {
+            return new ByteArrayInputStream(content);
+        }
+
+        @Override
+        public long getFileSize(String path) {
+            return content.length;
+        }
+    }
+
+    private static class FailingFileStorageService implements FileStorageService {
+
+        @Override
+        public String save(org.springframework.web.multipart.MultipartFile file) {
+            return null;
+        }
+
+        @Override
+        public void delete(String path) {
+        }
+
+        @Override
+        public InputStream readStream(String path) throws IOException {
+            throw new IOException("should not read remote url from storage");
+        }
+
+        @Override
+        public long getFileSize(String path) {
+            return 0L;
+        }
+    }
+}
--- a/easyflow-modules/easyflow-module-ai/src/test/java/tech/easyflow/ai/node/DocNodeFileContentExtractorTest.java
+++ b/easyflow-modules/easyflow-module-ai/src/test/java/tech/easyflow/ai/node/DocNodeFileContentExtractorTest.java
@@ -0,0 +1,278 @@
+package tech.easyflow.ai.node;
+
+import org.junit.Assert;
+import org.junit.Test;
+import tech.easyflow.ai.document.model.DocumentParseTaskInfo;
+import tech.easyflow.ai.document.model.DocumentParseTaskStatus;
+import tech.easyflow.ai.document.model.DocumentParsedResult;
+import tech.easyflow.ai.document.model.DocumentSourceRef;
+import tech.easyflow.ai.document.service.DocumentParseBridgeService;
+import tech.easyflow.common.filestorage.FileStorageService;
+import tech.easyflow.common.web.exceptions.BusinessException;
+
+import java.io.ByteArrayInputStream;
+import java.io.File;
+import java.io.IOException;
+import java.io.InputStream;
+import java.net.InetSocketAddress;
+import java.nio.charset.StandardCharsets;
+import java.util.HashMap;
+import java.util.Map;
+import com.sun.net.httpserver.HttpServer;
+
+/**
+ * {@link DocNodeFileContentExtractor} 单元测试。
+ *
+ * @author Codex
+ * @since 2026-04-14
+ */
+public class DocNodeFileContentExtractorTest {
+
+    /**
+     * 验证 PDF 文件会走统一文档解析桥接服务。
+     */
+    @Test
+    public void shouldUseDocumentBridgeForPdf() {
+        RecordingDocumentParseBridgeService bridgeService = new RecordingDocumentParseBridgeService();
+        DocNodeFileContentExtractor extractor = new DocNodeFileContentExtractor(
+            bridgeService,
+            new FakeFileStorageService(),
+            new FakeReaderManager("ignored")
+        );
+
+        String content = extractor.extract(buildFileValue("demo.pdf", "/files/demo.pdf", "application/pdf"));
+
+        Assert.assertEquals("# parsed", content);
+        Assert.assertNotNull(bridgeService.lastSource);
+        Assert.assertEquals("demo.pdf", bridgeService.lastSource.getFileName());
+    }
+
+    /**
+     * 验证非 PDF 文件会继续走默认读取器。
+     */
+    @Test
+    public void shouldUseDefaultReaderForNonPdf() {
+        RecordingDocumentParseBridgeService bridgeService = new RecordingDocumentParseBridgeService();
+        DocNodeFileContentExtractor extractor = new DocNodeFileContentExtractor(
+            bridgeService,
+            new FakeFileStorageService(),
+            new FakeReaderManager("plain text")
+        );
+
+        String content = extractor.extract(buildFileValue("demo.docx", "/files/demo.docx", "application/vnd.openxmlformats-officedocument.wordprocessingml.document"));
+
+        Assert.assertEquals("plain text", content);
+        Assert.assertNull(bridgeService.lastSource);
+    }
+
+    /**
+     * 验证缺少 filePath 时会抛出明确异常。
+     */
+    @Test
+    public void shouldRejectMissingFilePath() {
+        DocNodeFileContentExtractor extractor = new DocNodeFileContentExtractor(
+            new RecordingDocumentParseBridgeService(),
+            new FakeFileStorageService(),
+            new FakeReaderManager("plain text")
+        );
+
+        try {
+            extractor.extract(buildFileValue("demo.pdf", null, "application/pdf"));
+            Assert.fail("expected BusinessException");
+        } catch (BusinessException e) {
+            Assert.assertEquals("文件输入缺少 filePath", e.getMessage());
+        }
+    }
+
+    /**
+     * 验证解析结果为空时不会回退旧 PDF 读取链路。
+     */
+    @Test
+    public void shouldFailWhenPdfParseResultIsEmpty() {
+        RecordingDocumentParseBridgeService bridgeService = new RecordingDocumentParseBridgeService();
+        bridgeService.response.setPreferredText(null);
+        bridgeService.response.setMarkdown(null);
+        bridgeService.response.setPlainText(null);
+        DocNodeFileContentExtractor extractor = new DocNodeFileContentExtractor(
+            bridgeService,
+            new FakeFileStorageService(),
+            new FakeReaderManager("pdf fallback")
+        );
+
+        try {
+            extractor.extract(buildFileValue("demo.pdf", "/files/demo.pdf", "application/pdf"));
+            Assert.fail("expected BusinessException");
+        } catch (BusinessException e) {
+            Assert.assertEquals("PDF 文档解析结果为空", e.getMessage());
+        }
+    }
+
+    /**
+     * 验证远端素材 URL 的非 PDF 文件不会误走本地存储读取。
+     */
+    @Test
+    public void shouldReadRemoteUrlForNonPdf() {
+        RecordingDocumentParseBridgeService bridgeService = new RecordingDocumentParseBridgeService();
+        HttpServer server;
+        try {
+            server = HttpServer.create(new InetSocketAddress(0), 0);
+        } catch (IOException e) {
+            throw new RuntimeException(e);
+        }
+        byte[] body = "remote text".getBytes(StandardCharsets.UTF_8);
+        server.createContext("/demo.docx", exchange -> {
+            exchange.sendResponseHeaders(200, body.length);
+            exchange.getResponseBody().write(body);
+            exchange.close();
+        });
+        server.start();
+        try {
+            DocNodeFileContentExtractor extractor = new DocNodeFileContentExtractor(
+                bridgeService,
+                new FailingFileStorageService(),
+                new ReadingReaderManager()
+            );
+
+            String content = extractor.extract(buildFileValue(
+                "demo.docx",
+                "http://127.0.0.1:" + server.getAddress().getPort() + "/demo.docx",
+                ""
+            ));
+
+            Assert.assertEquals("remote text", content);
+            Assert.assertNull(bridgeService.lastSource);
+        } finally {
+            server.stop(0);
+        }
+    }
+
+    private Map<String, Object> buildFileValue(String fileName, String filePath, String contentType) {
+        Map<String, Object> value = new HashMap<String, Object>();
+        value.put("fileName", fileName);
+        value.put("filePath", filePath);
+        value.put("contentType", contentType);
+        value.put("size", 16L);
+        value.put("url", filePath);
+        return value;
+    }
+
+    private static class RecordingDocumentParseBridgeService implements DocumentParseBridgeService {
+
+        private final DocumentParsedResult response = new DocumentParsedResult();
+        private DocumentSourceRef lastSource;
+
+        private RecordingDocumentParseBridgeService() {
+            response.setPreferredText("# parsed");
+            response.setMarkdown("# parsed");
+            response.setPlainText("parsed");
+        }
+
+        @Override
+        public DocumentParsedResult parse(DocumentSourceRef source, tech.easyflow.ai.document.model.DocumentParseScenario scenario) {
+            this.lastSource = source;
+            return response;
+        }
+
+        @Override
+        public DocumentParseTaskStatus submit(DocumentSourceRef source, tech.easyflow.ai.document.model.DocumentParseScenario scenario) {
+            return new DocumentParseTaskStatus();
+        }
+
+        @Override
+        public DocumentParseTaskStatus queryTask(String taskId) {
+            return new DocumentParseTaskStatus();
+        }
+
+        @Override
+        public DocumentParsedResult queryResult(String taskId) {
+            return response;
+        }
+
+        @Override
+        public DocumentParseTaskInfo queryTaskInfo(String taskId) {
+            return new DocumentParseTaskInfo();
+        }
+    }
+
+    private static class FakeFileStorageService implements FileStorageService {
+
+        @Override
+        public String save(org.springframework.web.multipart.MultipartFile file) {
+            return null;
+        }
+
+        @Override
+        public void delete(String path) {
+        }
+
+        @Override
+        public String save(File file, String prePath) {
+            return null;
+        }
+
+        @Override
+        public InputStream readStream(String path) throws IOException {
+            return new ByteArrayInputStream("doc".getBytes(StandardCharsets.UTF_8));
+        }
+
+        @Override
+        public long getFileSize(String path) {
+            return 3L;
+        }
+    }
+
+    private static class FakeReaderManager extends ReaderManager {
+
+        private final String content;
+
+        private FakeReaderManager(String content) {
+            this.content = content;
+        }
+
+        @Override
+        public ReadDocService getReader() {
+            return (fileName, is) -> content;
+        }
+    }
+
+    private static class ReadingReaderManager extends ReaderManager {
+
+        @Override
+        public ReadDocService getReader() {
+            return (fileName, is) -> {
+                try {
+                    return new String(is.readAllBytes(), StandardCharsets.UTF_8);
+                } catch (IOException e) {
+                    throw new RuntimeException(e);
+                }
+            };
+        }
+    }
+
+    private static class FailingFileStorageService implements FileStorageService {
+
+        @Override
+        public String save(org.springframework.web.multipart.MultipartFile file) {
+            return null;
+        }
+
+        @Override
+        public void delete(String path) {
+        }
+
+        @Override
+        public String save(File file, String prePath) {
+            return null;
+        }
+
+        @Override
+        public InputStream readStream(String path) throws IOException {
+            throw new IOException("should not read remote url from storage");
+        }
+
+        @Override
+        public long getFileSize(String path) {
+            return 0L;
+        }
+    }
+}
--- a/easyflow-starter/easyflow-starter-all/src/main/resources/application.yml
+++ b/easyflow-starter/easyflow-starter-all/src/main/resources/application.yml
@@ -105,7 +105,7 @@ easyflow:
  analytical-db:
    # 是否启用分析数据库
    enabled: true
-    url: jdbc:clickhouse://127.0.0.1:38123/easyflow?jdbc_ignore_unsupported_values=true&socket_timeout=30000&compress=false&ssl=false
+    url: jdbc:clickhouse://114.66.22.180:38123/easyflow_dev?jdbc_ignore_unsupported_values=true&socket_timeout=30000&compress=false&ssl=false
    username: easyflow
    password: 123456
    driver-class-name: com.clickhouse.jdbc.ClickHouseDriver
@@ -163,6 +163,17 @@ dromara:
        domain: http://127.0.0.1:39000/easyflow/
        base-path: attachment

+# easy-agents 文档解析统一配置
+easy-agents:
+  document:
+    pdf:
+      provider: mineru
+      mineru:
+        # 统一文档解析桥接层直接复用 easy-agents 的 provider 配置，不在 easyflow 再复制一套配置体系
+        base-url: https://hub.wust.edu.cn/modelServer/mineru-api
+        default-lang-list:
+          - ch
+
 # 自定义节点相关配置
 node:
  # 文件内容提取节点，默认使用简单文档读取器，可自行实现 ReadDocService
--- a/easyflow-ui-admin/app/src/locales/langs/en-US/aiWorkflow.json
+++ b/easyflow-ui-admin/app/src/locales/langs/en-US/aiWorkflow.json
@@ -26,7 +26,8 @@
  "result": "Result",
  "confirm": "For contents to be confirmed, please confirm first!",
  "completed": "Chain has been completed, please start a new one.",
-  "fileContentExtraction": "FileContentExtraction",
+  "fileContentExtraction": "Document Parsing",
+  "documentInput": "FileInput",
  "documentAddress": "DocumentAddress",
  "parsedText": "ParsedText",
  "resourceSync": "ResourceSync",
@@ -94,7 +95,8 @@
  "stageSave": "Save Check",
  "stagePreExecute": "Pre-execute Check",
  "descriptions": {
-    "fileContentExtraction": "Extract text content from PDF or Word documents, etc",
+    "fileContentExtraction": "Parse content from PDF, Word, and other document files",
+    "documentInput": "Upload a file or choose one from resources",
    "documentAddress": "Document URL address",
    "parsedText": "Parsed text content",
    "resourceSync": "Download resource files and save to system resource library",
--- a/easyflow-ui-admin/app/src/locales/langs/zh-CN/aiWorkflow.json
+++ b/easyflow-ui-admin/app/src/locales/langs/zh-CN/aiWorkflow.json
@@ -26,7 +26,8 @@
  "result": "执行结果",
  "confirm": "有待确认的内容，请先确认！",
  "completed": "流程已执行完毕，请重新发起。",
-  "fileContentExtraction": "文件内容提取",
+  "fileContentExtraction": "文档解析",
+  "documentInput": "文件输入",
  "documentAddress": "文档地址",
  "parsedText": "解析后的文本",
  "resourceSync": "素材同步",
@@ -94,7 +95,8 @@
  "stageSave": "保存校验",
  "stagePreExecute": "预执行校验",
  "descriptions": {
-    "fileContentExtraction": "提取 PDF 或者 Word 等文件中的文字内容",
+    "fileContentExtraction": "解析 PDF 或 Word 等文档内容",
+    "documentInput": "上传文件或从素材库中选择文件",
    "documentAddress": "文档的url地址",
    "parsedText": "解析后的文本内容",
    "resourceSync": "下载素材文件并保存到系统素材库",
--- a/easyflow-ui-admin/app/src/views/ai/workflow/components/WorkflowFileInput.vue
+++ b/easyflow-ui-admin/app/src/views/ai/workflow/components/WorkflowFileInput.vue
@@ -0,0 +1,161 @@
+<script setup lang="ts">
+import { computed, ref } from 'vue';
+
+import { ElButton, ElLink, ElMessage } from 'element-plus';
+
+import { api } from '#/api/request';
+import { $t } from '#/locales';
+import ChooseResource from '#/views/ai/resource/ChooseResource.vue';
+
+import {
+  buildWorkflowFileValueFromResource,
+  buildWorkflowFileValueFromUpload,
+  formatWorkflowFileSize,
+  isWorkflowFileValue,
+} from './workflowFileValue';
+
+const props = defineProps({
+  modelValue: {
+    type: Object as () => Record<string, any> | undefined,
+    default: undefined,
+  },
+});
+
+const emit = defineEmits(['update:modelValue']);
+
+const uploadLoading = ref(false);
+const fileInputRef = ref<HTMLInputElement | null>(null);
+
+const currentFile = computed(() =>
+  isWorkflowFileValue(props.modelValue) ? props.modelValue : undefined,
+);
+
+function triggerSelectFile() {
+  if (uploadLoading.value) {
+    return;
+  }
+  fileInputRef.value?.click();
+}
+
+async function handleNativeFileChange(event: Event) {
+  const input = event.target as HTMLInputElement;
+  const file = input.files?.[0];
+  if (!file) {
+    return;
+  }
+  uploadLoading.value = true;
+  try {
+    const res = await api.upload('/api/v1/commons/upload', { file }, {});
+    const fileValue = buildWorkflowFileValueFromUpload(file, res?.data?.path);
+    emit('update:modelValue', fileValue);
+  } catch (error) {
+    ElMessage.error('文件上传失败');
+    console.error('工作流文件上传失败', error);
+  } finally {
+    uploadLoading.value = false;
+    input.value = '';
+  }
+}
+
+function handleChooseResource(resource: any) {
+  try {
+    const fileValue = buildWorkflowFileValueFromResource(resource || {});
+    emit('update:modelValue', fileValue);
+  } catch (error: any) {
+    ElMessage.error(error?.message || '素材文件选择失败');
+  }
+}
+
+function clearFile() {
+  emit('update:modelValue', undefined);
+}
+</script>
+
+<template>
+  <div class="workflow-file-input">
+    <input
+      ref="fileInputRef"
+      class="workflow-file-input__native"
+      type="file"
+      @change="handleNativeFileChange"
+    />
+
+    <div v-if="currentFile" class="workflow-file-input__summary">
+      <div class="workflow-file-input__name">
+        {{ currentFile.fileName }}
+      </div>
+      <div class="workflow-file-input__meta">
+        <span>{{ formatWorkflowFileSize(currentFile.size) }}</span>
+        <ElLink
+          v-if="currentFile.url || currentFile.filePath"
+          :href="currentFile.url || currentFile.filePath"
+          target="_blank"
+          type="primary"
+        >
+          {{ $t('button.view') }}
+        </ElLink>
+      </div>
+    </div>
+
+    <div class="workflow-file-input__actions">
+      <ElButton
+        type="primary"
+        plain
+        :loading="uploadLoading"
+        @click="triggerSelectFile"
+      >
+        {{ currentFile ? $t('button.replace') : $t('button.upload') }}
+      </ElButton>
+      <ChooseResource attr-name="file" @choose="handleChooseResource" />
+      <ElButton
+        v-if="currentFile"
+        text
+        type="danger"
+        @click="clearFile"
+      >
+        {{ $t('button.delete') }}
+      </ElButton>
+    </div>
+  </div>
+</template>
+
+<style scoped>
+.workflow-file-input {
+  display: flex;
+  flex-direction: column;
+  gap: 10px;
+}
+
+.workflow-file-input__native {
+  display: none;
+}
+
+.workflow-file-input__summary {
+  padding: 12px;
+  border: 1px solid var(--el-border-color-light);
+  border-radius: 10px;
+  background: var(--el-fill-color-blank);
+}
+
+.workflow-file-input__name {
+  font-size: 13px;
+  font-weight: 600;
+  color: var(--el-text-color-primary);
+  word-break: break-word;
+}
+
+.workflow-file-input__meta {
+  display: flex;
+  align-items: center;
+  gap: 10px;
+  margin-top: 6px;
+  font-size: 12px;
+  color: var(--el-text-color-secondary);
+}
+
+.workflow-file-input__actions {
+  display: flex;
+  flex-wrap: wrap;
+  gap: 8px;
+}
+</style>
--- a/easyflow-ui-admin/app/src/views/ai/workflow/components/WorkflowFormItem.vue
+++ b/easyflow-ui-admin/app/src/views/ai/workflow/components/WorkflowFormItem.vue
@@ -10,6 +10,7 @@ import {

 import { $t } from '#/locales';
 import ChooseResource from '#/views/ai/resource/ChooseResource.vue';
+import WorkflowFileInput from '#/views/ai/workflow/components/WorkflowFileInput.vue';

 const props = defineProps({
  parameters: {
@@ -27,10 +28,19 @@ const props = defineProps({
 });
 const emit = defineEmits(['update:runParams']);
 function getContentType(item: any) {
-  return item.contentType || 'text';
+  if (item.contentType) {
+    return item.contentType;
+  }
+  if (String(item.dataType || '').toLowerCase() === 'file') {
+    return 'file';
+  }
+  return 'text';
 }
 function isResource(contentType: any) {
-  return ['audio', 'file', 'image', 'video'].includes(contentType);
+  return ['audio', 'image', 'video'].includes(contentType);
+}
+function isFileContentType(contentType: any) {
+  return contentType === 'file';
 }
 function getCheckboxOptions(item: any) {
  if (item.enums) {
@@ -105,6 +115,12 @@ function choose(data: any, propName: string) {
        :placeholder="item.formPlaceholder"
      />
    </template>
+    <template v-if="isFileContentType(getContentType(item))">
+      <WorkflowFileInput
+        :model-value="runParams[item.name]"
+        @update:model-value="(val) => updateParam(item.name, val)"
+      />
+    </template>
    <template v-if="isResource(getContentType(item))">
      <ElInput
        :model-value="runParams[item.name]"
--- a/easyflow-ui-admin/app/src/views/ai/workflow/components/tests/workflowFileValue.test.ts
+++ b/easyflow-ui-admin/app/src/views/ai/workflow/components/tests/workflowFileValue.test.ts
@@ -0,0 +1,52 @@
+import { describe, expect, it } from 'vitest';
+
+import {
+  buildWorkflowFileValueFromResource,
+  buildWorkflowFileValueFromUpload,
+  formatWorkflowFileSize,
+  isWorkflowFileValue,
+} from '../workflowFileValue';
+
+describe('workflowFileValue', () => {
+  it('从上传结果构建统一文件对象', () => {
+    const file = new File(['demo'], 'demo.pdf', { type: 'application/pdf' });
+
+    const value = buildWorkflowFileValueFromUpload(file, 'https://example.com/demo.pdf');
+
+    expect(value).toEqual({
+      fileName: 'demo.pdf',
+      filePath: 'https://example.com/demo.pdf',
+      contentType: 'application/pdf',
+      size: 4,
+      url: 'https://example.com/demo.pdf',
+    });
+  });
+
+  it('从素材对象构建统一文件对象', () => {
+    const value = buildWorkflowFileValueFromResource({
+      fileSize: '128',
+      resourceName: 'manual',
+      resourceUrl: 'https://example.com/manual.docx',
+      suffix: 'docx',
+    });
+
+    expect(value).toEqual({
+      fileName: 'manual.docx',
+      filePath: 'https://example.com/manual.docx',
+      contentType: '',
+      size: 128,
+      url: 'https://example.com/manual.docx',
+    });
+  });
+
+  it('正确识别有效文件值并格式化大小', () => {
+    expect(
+      isWorkflowFileValue({
+        fileName: 'demo.pdf',
+        filePath: 'https://example.com/demo.pdf',
+      }),
+    ).toBe(true);
+    expect(isWorkflowFileValue({ fileName: 'demo.pdf' })).toBe(false);
+    expect(formatWorkflowFileSize(2048)).toBe('2.0 KB');
+  });
+});
--- a/easyflow-ui-admin/app/src/views/ai/workflow/components/workflowFileValue.ts
+++ b/easyflow-ui-admin/app/src/views/ai/workflow/components/workflowFileValue.ts
@@ -0,0 +1,109 @@
+/**
+ * 工作流运行态单文件值。
+ */
+export interface WorkflowFileValue {
+  fileName: string;
+  filePath: string;
+  contentType?: string;
+  size?: number;
+  url?: string;
+}
+
+/**
+ * 素材对象最小字段约束。
+ */
+export interface WorkflowResourceLike {
+  fileSize?: number | string;
+  resourceName?: string;
+  resourceUrl?: string;
+  suffix?: string;
+}
+
+/**
+ * 基于上传结果构建工作流文件值。
+ */
+export function buildWorkflowFileValueFromUpload(
+  file: File,
+  path: string,
+): WorkflowFileValue {
+  const resolvedPath = String(path || '').trim();
+  if (!resolvedPath) {
+    throw new Error('上传结果缺少文件路径');
+  }
+  return {
+    fileName: file.name,
+    filePath: resolvedPath,
+    contentType: file.type || '',
+    size: file.size,
+    url: resolvedPath,
+  };
+}
+
+/**
+ * 基于素材对象构建工作流文件值。
+ */
+export function buildWorkflowFileValueFromResource(
+  resource: WorkflowResourceLike,
+): WorkflowFileValue {
+  const resourceUrl = String(resource?.resourceUrl || '').trim();
+  if (!resourceUrl) {
+    throw new Error('素材缺少 resourceUrl');
+  }
+  const resourceName = String(resource?.resourceName || '').trim();
+  const suffix = String(resource?.suffix || '').trim();
+  const fallbackFileName = resourceUrl.split('/').pop()?.split('?')[0] || '';
+  return {
+    fileName:
+      resourceName && suffix
+        ? `${resourceName}.${suffix}`
+        : resourceName || fallbackFileName,
+    filePath: resourceUrl,
+    contentType: '',
+    size: toNumber(resource?.fileSize),
+    url: resourceUrl,
+  };
+}
+
+/**
+ * 判断是否为有效的工作流文件值。
+ */
+export function isWorkflowFileValue(value: unknown): value is WorkflowFileValue {
+  if (!value || typeof value !== 'object') {
+    return false;
+  }
+  const candidate = value as Partial<WorkflowFileValue>;
+  return Boolean(
+    String(candidate.fileName || '').trim() &&
+      String(candidate.filePath || '').trim(),
+  );
+}
+
+/**
+ * 友好格式化文件大小。
+ */
+export function formatWorkflowFileSize(size?: number): string {
+  if (!size || Number.isNaN(size) || size <= 0) {
+    return '-';
+  }
+  if (size < 1024) {
+    return `${size} B`;
+  }
+  if (size < 1024 * 1024) {
+    return `${(size / 1024).toFixed(1)} KB`;
+  }
+  if (size < 1024 * 1024 * 1024) {
+    return `${(size / (1024 * 1024)).toFixed(1)} MB`;
+  }
+  return `${(size / (1024 * 1024 * 1024)).toFixed(1)} GB`;
+}
+
+function toNumber(value?: number | string) {
+  if (typeof value === 'number') {
+    return value;
+  }
+  if (typeof value === 'string' && value.trim()) {
+    const parsed = Number(value);
+    return Number.isFinite(parsed) ? parsed : undefined;
+  }
+  return undefined;
+}
--- a/easyflow-ui-admin/app/src/views/ai/workflow/customNode/documentNode.ts
+++ b/easyflow-ui-admin/app/src/views/ai/workflow/customNode/documentNode.ts
@@ -13,12 +13,13 @@ export default {
    outputDefsAddEnable: false,
    parameters: [
      {
-        name: 'fileUrl',
+        name: 'file',
        nameDisabled: true,
-        title: $t('aiWorkflow.documentAddress'),
+        title: $t('aiWorkflow.documentInput'),
        dataType: 'File',
+        contentType: 'file',
        required: true,
-        description: $t('aiWorkflow.descriptions.documentAddress'),
+        description: $t('aiWorkflow.descriptions.documentInput'),
      },
    ],
    outputDefs: [
--- a/pom.xml
+++ b/pom.xml
@@ -182,6 +182,11 @@
                <artifactId>easy-agents-support</artifactId>
                <version>${easy-agents.version}</version>
            </dependency>
+            <dependency>
+                <groupId>com.easyagents</groupId>
+                <artifactId>easy-agents-document-core</artifactId>
+                <version>${easy-agents.version}</version>
+            </dependency>
            <dependency>
                <groupId>com.easyagents</groupId>
                <artifactId>easy-agents-rag-retrieval</artifactId>