After updating the Document in spring ai core, I encountered the following problem

This is my code：

   Resource resource = new FileSystemResource(filePath);
    List<Document> documents = new TikaDocumentReader(resource).read();
    return new TokenTextSplitter(knowledgeBaseFileSlice.getDefaultChunkSize(), 
            knowledgeBaseFileSlice.getMinChunkSizeChars(),
            knowledgeBaseFileSlice.getMinChunkLengthToEmbed(), knowledgeBaseFileSlice.getMaxNumChunks(),
            knowledgeBaseFileSlice.isKeepSeparator()).apply(documents);

After TikaDocumentReader reads a Word document, the content read not only includes the text of the document, but also the XML information of the file, if I use getText(), the output will include the following content, like this：
docProps/app.xml
Normal.dotm 1 0 0 0 0 0 false false 0 WPS Office_10.1.0.7698_F1E327BC-269C-435d-A152-05C5408002CA 0

docProps/core.xml
2023-08-26T18:18:00Z admin admin 2023-08-26T18:18:41Z 1

docProps/custom.xml
2052-10.1.0.7698

word/styles.xml

word/settings.xml

word/theme/theme1.xml

word/document.xml

What went wrong？？？

Start by reproducing the behavior with TikaDocumentReader on the reported Word document and inspect the value returned by getText() before TokenTextSplitter is applied. Compare the extracted text with the document contents and determine whether the XML entries are expected; done means the cause and an appropriate extraction behavior are clearly established.