feat: 支持工作流文档节点解析 Office 新格式
- DocNode 对 pdf/docx/pptx/xlsx 统一走桥接同步解析 - 修复 DOCX 误回退到 PDF 解析服务的问题并补齐回归测试
This commit is contained in:
@@ -101,7 +101,7 @@ public class DocumentParseBridgeServiceImplTest {
|
||||
@Test
|
||||
public void shouldRoutePptxToDedicatedService() {
|
||||
FakePptxDocumentParseService pptxService = new FakePptxDocumentParseService();
|
||||
FakePdfDocumentParseService defaultService = new FakePdfDocumentParseService();
|
||||
FakeDefaultDocumentParseService defaultService = new FakeDefaultDocumentParseService();
|
||||
DocumentParseBridgeServiceImpl bridgeService = buildBridgeService(null, pptxService, null, defaultService);
|
||||
|
||||
DocumentParsedResult result = bridgeService.parse(buildSource("slides.pptx",
|
||||
@@ -112,6 +112,61 @@ public class DocumentParseBridgeServiceImplTest {
|
||||
Assert.assertEquals(0, defaultService.parseCallCount);
|
||||
}
|
||||
|
||||
/**
|
||||
* 验证 DOCX 会路由到默认桥接服务，而不是误走 PDF 服务。
|
||||
*/
|
||||
@Test
|
||||
public void shouldRouteDocxToDefaultBridgeService() {
|
||||
FakePdfDocumentParseService pdfService = new FakePdfDocumentParseService();
|
||||
FakeDefaultDocumentParseService defaultService = new FakeDefaultDocumentParseService();
|
||||
DocumentParseBridgeServiceImpl bridgeService = buildBridgeService(pdfService, null, null, defaultService);
|
||||
|
||||
DocumentParsedResult result = bridgeService.parse(buildSource("demo.docx",
|
||||
"application/vnd.openxmlformats-officedocument.wordprocessingml.document"), DocumentParseScenario.WORKFLOW_TEXT);
|
||||
|
||||
Assert.assertEquals("# docx", result.getPreferredText());
|
||||
Assert.assertEquals(1, defaultService.parseCallCount);
|
||||
Assert.assertEquals(0, pdfService.parseCallCount);
|
||||
}
|
||||
|
||||
/**
|
||||
* 验证 DOCX 在缺少默认桥接服务时会明确失败，而不是退到 PDF 服务。
|
||||
*/
|
||||
@Test
|
||||
public void shouldThrowWhenDocxBridgeServiceDisabled() {
|
||||
FakePdfDocumentParseService pdfService = new FakePdfDocumentParseService();
|
||||
DocumentParseBridgeServiceImpl bridgeService = buildBridgeService(pdfService, null, null, null);
|
||||
|
||||
try {
|
||||
bridgeService.parse(buildSource("demo.docx",
|
||||
"application/vnd.openxmlformats-officedocument.wordprocessingml.document"), DocumentParseScenario.WORKFLOW_TEXT);
|
||||
Assert.fail("expected DocumentParseBridgeException");
|
||||
} catch (DocumentParseBridgeException e) {
|
||||
Assert.assertEquals("service_not_enabled", e.getCode());
|
||||
Assert.assertTrue(e.getMessage().contains("DOCX"));
|
||||
}
|
||||
Assert.assertEquals(0, pdfService.parseCallCount);
|
||||
}
|
||||
|
||||
/**
|
||||
* 验证 XLSX 在缺少专用桥接服务时会明确失败。
|
||||
*/
|
||||
@Test
|
||||
public void shouldThrowWhenXlsxBridgeServiceDisabled() {
|
||||
FakeDefaultDocumentParseService defaultService = new FakeDefaultDocumentParseService();
|
||||
DocumentParseBridgeServiceImpl bridgeService = buildBridgeService(null, null, null, defaultService);
|
||||
|
||||
try {
|
||||
bridgeService.parse(buildSource("table.xlsx",
|
||||
"application/vnd.openxmlformats-officedocument.spreadsheetml.sheet"), DocumentParseScenario.WORKFLOW_TEXT);
|
||||
Assert.fail("expected DocumentParseBridgeException");
|
||||
} catch (DocumentParseBridgeException e) {
|
||||
Assert.assertEquals("service_not_enabled", e.getCode());
|
||||
Assert.assertTrue(e.getMessage().contains("XLSX"));
|
||||
}
|
||||
Assert.assertEquals(0, defaultService.parseCallCount);
|
||||
}
|
||||
|
||||
private DocumentParseBridgeServiceImpl buildBridgeService(PdfDocumentParseService pdfDocumentParseService,
|
||||
PptxDocumentParseService pptxDocumentParseService,
|
||||
XlsxDocumentParseService xlsxDocumentParseService,
|
||||
@@ -217,6 +272,45 @@ public class DocumentParseBridgeServiceImplTest {
|
||||
}
|
||||
}
|
||||
|
||||
private static class FakeDefaultDocumentParseService implements DocumentParseService<ParseRequest> {
|
||||
|
||||
private ParseRequest lastParseRequest;
|
||||
private int parseCallCount;
|
||||
|
||||
@Override
|
||||
public ParseResponse parse(ParseRequest request) {
|
||||
parseCallCount++;
|
||||
lastParseRequest = request;
|
||||
ParseResult result = new ParseResult();
|
||||
result.setFileName("demo.docx");
|
||||
result.setMarkdown("# docx");
|
||||
result.setPlainText("docx");
|
||||
ParseResponse response = new ParseResponse();
|
||||
response.setResults(Collections.singletonList(result));
|
||||
return response;
|
||||
}
|
||||
|
||||
@Override
|
||||
public ParseTaskStatus submit(ParseRequest request) {
|
||||
throw new UnsupportedOperationException();
|
||||
}
|
||||
|
||||
@Override
|
||||
public ParseTaskStatus queryTask(String taskId) {
|
||||
throw new UnsupportedOperationException();
|
||||
}
|
||||
|
||||
@Override
|
||||
public ParseResponse queryResult(String taskId) {
|
||||
throw new UnsupportedOperationException();
|
||||
}
|
||||
|
||||
@Override
|
||||
public ParseTaskInfo queryTaskInfo(String taskId) {
|
||||
throw new UnsupportedOperationException();
|
||||
}
|
||||
}
|
||||
|
||||
private static class FakePptxDocumentParseService implements PptxDocumentParseService {
|
||||
|
||||
private int parseCallCount;
|
||||
|
||||
@@ -48,10 +48,75 @@ public class DocNodeFileContentExtractorTest {
|
||||
}
|
||||
|
||||
/**
|
||||
* 验证非 PDF 文件会继续走默认读取器。
|
||||
* 验证 DOCX 文件会走统一文档解析桥接服务。
|
||||
*/
|
||||
@Test
|
||||
public void shouldUseDefaultReaderForNonPdf() {
|
||||
public void shouldUseDocumentBridgeForDocx() {
|
||||
RecordingDocumentParseBridgeService bridgeService = new RecordingDocumentParseBridgeService();
|
||||
DocNodeFileContentExtractor extractor = new DocNodeFileContentExtractor(
|
||||
bridgeService,
|
||||
new FakeFileStorageService(),
|
||||
new FakeReaderManager("ignored")
|
||||
);
|
||||
|
||||
String content = extractor.extract(buildFileValue("demo.docx", "/files/demo.docx", "application/vnd.openxmlformats-officedocument.wordprocessingml.document"));
|
||||
|
||||
Assert.assertEquals("# parsed", content);
|
||||
Assert.assertNotNull(bridgeService.lastSource);
|
||||
Assert.assertEquals("demo.docx", bridgeService.lastSource.getFileName());
|
||||
}
|
||||
|
||||
/**
|
||||
* 验证 PPTX 文件会走统一文档解析桥接服务。
|
||||
*/
|
||||
@Test
|
||||
public void shouldUseDocumentBridgeForPptx() {
|
||||
RecordingDocumentParseBridgeService bridgeService = new RecordingDocumentParseBridgeService();
|
||||
DocNodeFileContentExtractor extractor = new DocNodeFileContentExtractor(
|
||||
bridgeService,
|
||||
new FakeFileStorageService(),
|
||||
new FakeReaderManager("ignored")
|
||||
);
|
||||
|
||||
String content = extractor.extract(buildFileValue(
|
||||
"slides.pptx",
|
||||
"/files/slides.pptx",
|
||||
"application/vnd.openxmlformats-officedocument.presentationml.presentation"
|
||||
));
|
||||
|
||||
Assert.assertEquals("# parsed", content);
|
||||
Assert.assertNotNull(bridgeService.lastSource);
|
||||
Assert.assertEquals("slides.pptx", bridgeService.lastSource.getFileName());
|
||||
}
|
||||
|
||||
/**
|
||||
* 验证 XLSX 文件会走统一文档解析桥接服务。
|
||||
*/
|
||||
@Test
|
||||
public void shouldUseDocumentBridgeForXlsx() {
|
||||
RecordingDocumentParseBridgeService bridgeService = new RecordingDocumentParseBridgeService();
|
||||
DocNodeFileContentExtractor extractor = new DocNodeFileContentExtractor(
|
||||
bridgeService,
|
||||
new FakeFileStorageService(),
|
||||
new FakeReaderManager("ignored")
|
||||
);
|
||||
|
||||
String content = extractor.extract(buildFileValue(
|
||||
"table.xlsx",
|
||||
"/files/table.xlsx",
|
||||
"application/vnd.openxmlformats-officedocument.spreadsheetml.sheet"
|
||||
));
|
||||
|
||||
Assert.assertEquals("# parsed", content);
|
||||
Assert.assertNotNull(bridgeService.lastSource);
|
||||
Assert.assertEquals("table.xlsx", bridgeService.lastSource.getFileName());
|
||||
}
|
||||
|
||||
/**
|
||||
* 验证非桥接类型文件会继续走默认读取器。
|
||||
*/
|
||||
@Test
|
||||
public void shouldUseDefaultReaderForUnsupportedType() {
|
||||
RecordingDocumentParseBridgeService bridgeService = new RecordingDocumentParseBridgeService();
|
||||
DocNodeFileContentExtractor extractor = new DocNodeFileContentExtractor(
|
||||
bridgeService,
|
||||
@@ -59,7 +124,7 @@ public class DocNodeFileContentExtractorTest {
|
||||
new FakeReaderManager("plain text")
|
||||
);
|
||||
|
||||
String content = extractor.extract(buildFileValue("demo.docx", "/files/demo.docx", "application/vnd.openxmlformats-officedocument.wordprocessingml.document"));
|
||||
String content = extractor.extract(buildFileValue("note.txt", "/files/note.txt", "text/plain"));
|
||||
|
||||
Assert.assertEquals("plain text", content);
|
||||
Assert.assertNull(bridgeService.lastSource);
|
||||
@@ -85,10 +150,10 @@ public class DocNodeFileContentExtractorTest {
|
||||
}
|
||||
|
||||
/**
|
||||
* 验证解析结果为空时不会回退旧 PDF 读取链路。
|
||||
* 验证桥接解析结果为空时不会回退旧读取链路。
|
||||
*/
|
||||
@Test
|
||||
public void shouldFailWhenPdfParseResultIsEmpty() {
|
||||
public void shouldFailWhenBridgeParseResultIsEmpty() {
|
||||
RecordingDocumentParseBridgeService bridgeService = new RecordingDocumentParseBridgeService();
|
||||
bridgeService.response.setPreferredText(null);
|
||||
bridgeService.response.setMarkdown(null);
|
||||
@@ -103,15 +168,15 @@ public class DocNodeFileContentExtractorTest {
|
||||
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());
|
||||
Assert.assertEquals("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。
|
||||
* 验证远端素材 URL 的非桥接文件不会误走本地存储读取。
|
||||
*/
|
||||
@Test
|
||||
public void shouldReadRemoteUrlForNonPdf() {
|
||||
public void shouldReadRemoteUrlForUnsupportedType() {
|
||||
RecordingDocumentParseBridgeService bridgeService = new RecordingDocumentParseBridgeService();
|
||||
HttpServer server;
|
||||
try {
|
||||
@@ -120,7 +185,7 @@ public class DocNodeFileContentExtractorTest {
|
||||
throw new RuntimeException(e);
|
||||
}
|
||||
byte[] body = "remote text".getBytes(StandardCharsets.UTF_8);
|
||||
server.createContext("/demo.docx", exchange -> {
|
||||
server.createContext("/note.txt", exchange -> {
|
||||
exchange.sendResponseHeaders(200, body.length);
|
||||
exchange.getResponseBody().write(body);
|
||||
exchange.close();
|
||||
@@ -134,9 +199,9 @@ public class DocNodeFileContentExtractorTest {
|
||||
);
|
||||
|
||||
String content = extractor.extract(buildFileValue(
|
||||
"demo.docx",
|
||||
"http://127.0.0.1:" + server.getAddress().getPort() + "/demo.docx",
|
||||
""
|
||||
"note.txt",
|
||||
"http://127.0.0.1:" + server.getAddress().getPort() + "/note.txt",
|
||||
"text/plain"
|
||||
));
|
||||
|
||||
Assert.assertEquals("remote text", content);
|
||||
|
||||
Reference in New Issue
Block a user