package jnpf.fileStorage.dms;
|
|
import jnpf.exception.DataException;
|
import org.springframework.stereotype.Service;
|
import org.springframework.web.multipart.MultipartFile;
|
import org.w3c.dom.*;
|
|
import javax.xml.XMLConstants;
|
import javax.xml.parsers.DocumentBuilderFactory;
|
import javax.xml.transform.TransformerFactory;
|
import javax.xml.transform.dom.DOMSource;
|
import javax.xml.transform.stream.StreamResult;
|
import java.io.*;
|
import java.net.URI;
|
import java.nio.file.Files;
|
import java.util.*;
|
import java.util.zip.*;
|
|
/** Cleans a new formal copy, rejecting revisions and leaving the source review attachment intact. */
|
@Service
|
public class DmsDocxCommentCleaner {
|
private static final int MAX_PARTS = 4096;
|
private static final int MAX_BYTES = 50 * 1024 * 1024;
|
private static final int MAX_XML = 16 * 1024 * 1024;
|
private static final String CT = "http://schemas.openxmlformats.org/package/2006/content-types";
|
private static final String REL = "http://schemas.openxmlformats.org/package/2006/relationships";
|
private static final Set<String> COMMENT_TYPES = Set.of("comments", "commentsExtended", "commentsIds",
|
"commentsExtensible", "people");
|
private static final Set<String> MARKERS = Set.of("commentRangeStart", "commentRangeEnd", "commentReference");
|
|
public MultipartFile clean(MultipartFile source) {
|
String name = source.getOriginalFilename();
|
if (name == null || !name.toLowerCase(Locale.ROOT).endsWith(".docx")) return source;
|
try {
|
Map<String, byte[]> parts = read(source);
|
Document types = parse(required(parts, "[Content_Types].xml"));
|
Map<String, Document> xml = new LinkedHashMap<>();
|
Set<String> removed = new HashSet<>();
|
boolean main = false;
|
NodeList overrides = types.getElementsByTagNameNS(CT, "Override");
|
for (int i = 0; i < overrides.getLength(); i++) {
|
Element type = (Element) overrides.item(i);
|
String contentType = type.getAttribute("ContentType");
|
String part = target("", type.getAttribute("PartName"));
|
if (commentContentType(contentType)) removed.add(part);
|
if (contentType.equals("application/vnd.openxmlformats-officedocument.wordprocessingml.document.main+xml")) {
|
required(parts, part);
|
main = true;
|
}
|
}
|
if (!main) throw new IOException("Missing DOCX main part");
|
// Parse every XML part securely. Relationship discovery also handles nonstandard comment part names.
|
for (Map.Entry<String, byte[]> part : parts.entrySet()) {
|
if (part.getKey().endsWith(".xml") || part.getKey().endsWith(".rels")) {
|
Document doc = parse(part.getValue());
|
xml.put(part.getKey(), doc);
|
if (part.getKey().endsWith(".rels")) {
|
NodeList links = doc.getElementsByTagNameNS(REL, "Relationship");
|
for (int i = 0; i < links.getLength(); i++) {
|
Element link = (Element) links.item(i);
|
if (commentRelationship(link.getAttribute("Type")) && !external(link))
|
removed.add(target(owner(part.getKey()), link.getAttribute("Target")));
|
}
|
}
|
}
|
}
|
// Remove comment-owned resources only if no retained part references them (shared images stay).
|
boolean expanded;
|
do {
|
expanded = false;
|
Set<String> candidates = new HashSet<>();
|
Set<String> retainedTargets = new HashSet<>();
|
for (Map.Entry<String, Document> part : xml.entrySet()) {
|
if (!part.getKey().endsWith(".rels")) continue;
|
String owner = owner(part.getKey());
|
NodeList links = part.getValue().getElementsByTagNameNS(REL, "Relationship");
|
for (int i = 0; i < links.getLength(); i++) {
|
Element link = (Element) links.item(i);
|
if (external(link)) continue;
|
String dest = target(owner, link.getAttribute("Target"));
|
if (removed.contains(owner)) candidates.add(dest);
|
else if (!commentRelationship(link.getAttribute("Type"))) retainedTargets.add(dest);
|
}
|
}
|
candidates.removeAll(retainedTargets);
|
if (removed.addAll(candidates)) expanded = true;
|
} while (expanded);
|
for (String part : new ArrayList<>(removed)) removed.add(rels(part));
|
boolean changed = parts.keySet().removeAll(removed);
|
for (Map.Entry<String, Document> part : xml.entrySet()) {
|
if (removed.contains(part.getKey())) continue;
|
Document doc = part.getValue();
|
boolean edited = DmsDocxRevisionRejector.reject(doc);
|
edited |= removeMarkers(doc);
|
if (part.getKey().endsWith(".rels")) {
|
NodeList links = doc.getElementsByTagNameNS(REL, "Relationship");
|
for (int i = links.getLength() - 1; i >= 0; i--) {
|
Element link = (Element) links.item(i);
|
if (commentRelationship(link.getAttribute("Type")) || !external(link)
|
&& removed.contains(target(owner(part.getKey()), link.getAttribute("Target")))) {
|
link.getParentNode().removeChild(link); edited = true;
|
}
|
}
|
}
|
if (part.getKey().equals("[Content_Types].xml")) {
|
NodeList entries = doc.getElementsByTagNameNS(CT, "Override");
|
for (int i = entries.getLength() - 1; i >= 0; i--) {
|
Element entry = (Element) entries.item(i);
|
if (removed.contains(target("", entry.getAttribute("PartName")))) {
|
entry.getParentNode().removeChild(entry); edited = true;
|
}
|
}
|
}
|
if (edited) { parts.put(part.getKey(), serialize(doc)); changed = true; }
|
}
|
if (!changed) return source;
|
// Package signatures cannot survive byte changes. Business signature-page content is unaffected.
|
if (parts.keySet().stream().anyMatch(p -> p.startsWith("_xmlsignatures/")))
|
throw new IOException("Digitally signed package cannot be rewritten");
|
ByteArrayOutputStream output = new ByteArrayOutputStream();
|
try (ZipOutputStream zip = new ZipOutputStream(output)) {
|
for (Map.Entry<String, byte[]> part : parts.entrySet()) {
|
zip.putNextEntry(new ZipEntry(part.getKey())); zip.write(part.getValue()); zip.closeEntry();
|
}
|
}
|
return new CleanedFile(source.getName(), name, source.getContentType(), output.toByteArray());
|
} catch (Exception e) {
|
throw new DataException("DOCX修订及批注清理失败，正式文件未存储：" + e.getMessage());
|
}
|
}
|
|
private static boolean commentContentType(String type) {
|
for (String name : COMMENT_TYPES) {
|
if (type.equals("application/vnd.openxmlformats-officedocument.wordprocessingml." + name + "+xml")
|
|| type.equals("application/vnd.ms-word." + name + "+xml")) return true;
|
}
|
return false;
|
}
|
private static boolean commentRelationship(String type) {
|
return (type.startsWith("http://schemas.openxmlformats.org/officeDocument/2006/relationships/")
|
|| type.startsWith("http://purl.oclc.org/ooxml/officeDocument/relationships/")
|
|| type.startsWith("http://schemas.microsoft.com/office/"))
|
&& COMMENT_TYPES.contains(type.substring(type.lastIndexOf('/') + 1));
|
}
|
private static boolean external(Element link) { return "External".equals(link.getAttribute("TargetMode")); }
|
private static String owner(String rels) throws IOException {
|
if (rels.equals("_rels/.rels")) return "";
|
int index = rels.lastIndexOf("/_rels/");
|
if (index < 0 || !rels.endsWith(".rels")) throw new IOException("Invalid relationship part");
|
return rels.substring(0, index + 1) + rels.substring(index + 7, rels.length() - 5);
|
}
|
private static String rels(String part) {
|
int index = part.lastIndexOf('/');
|
return part.substring(0, index + 1) + "_rels/" + part.substring(index + 1) + ".rels";
|
}
|
private static String target(String owner, String value) throws IOException {
|
try {
|
URI resolved = URI.create("/" + owner).resolve(value).normalize();
|
String path = resolved.getPath();
|
if (resolved.isAbsolute() || resolved.getAuthority() != null || path == null || !path.startsWith("/")
|
|| path.contains("/../")) throw new IOException("Invalid package target");
|
return path.substring(1);
|
} catch (IllegalArgumentException e) { throw new IOException("Invalid package target", e); }
|
}
|
private static byte[] required(Map<String, byte[]> parts, String name) throws IOException {
|
byte[] data = parts.get(name);
|
if (data == null) throw new IOException("Missing package part: " + name);
|
return data;
|
}
|
private static Map<String, byte[]> read(MultipartFile source) throws IOException {
|
Map<String, byte[]> parts = new LinkedHashMap<>();
|
int total = 0, count = 0;
|
try (ZipInputStream zip = new ZipInputStream(source.getInputStream())) {
|
ZipEntry entry;
|
while ((entry = zip.getNextEntry()) != null) {
|
String name = entry.getName();
|
if (++count > MAX_PARTS || name.startsWith("/") || name.contains("\\")
|
|| Arrays.asList(name.split("/")).contains("..") || parts.containsKey(name))
|
throw new IOException("Unsafe or oversized DOCX package");
|
if (entry.isDirectory()) continue;
|
ByteArrayOutputStream bytes = new ByteArrayOutputStream();
|
byte[] buffer = new byte[8192]; int read;
|
while ((read = zip.read(buffer)) != -1) {
|
total += read;
|
if (total > MAX_BYTES) throw new IOException("DOCX exceeds 50MB expanded limit");
|
bytes.write(buffer, 0, read);
|
}
|
parts.put(name, bytes.toByteArray());
|
}
|
}
|
return parts;
|
}
|
private static Document parse(byte[] bytes) throws Exception {
|
if (bytes.length > MAX_XML) throw new IOException("DOCX XML part exceeds 16MB limit");
|
DocumentBuilderFactory factory = DocumentBuilderFactory.newDefaultInstance(); factory.setNamespaceAware(true);
|
factory.setFeature(XMLConstants.FEATURE_SECURE_PROCESSING, true);
|
factory.setFeature("http://apache.org/xml/features/disallow-doctype-decl", true);
|
factory.setFeature("http://xml.org/sax/features/external-general-entities", false);
|
factory.setFeature("http://xml.org/sax/features/external-parameter-entities", false);
|
factory.setAttribute(XMLConstants.ACCESS_EXTERNAL_DTD, ""); factory.setAttribute(XMLConstants.ACCESS_EXTERNAL_SCHEMA, "");
|
return factory.newDocumentBuilder().parse(new ByteArrayInputStream(bytes));
|
}
|
private static boolean removeMarkers(Document doc) {
|
NodeList nodes = doc.getElementsByTagNameNS("*", "*"); boolean changed = false;
|
for (int i = nodes.getLength() - 1; i >= 0; i--) {
|
Node node = nodes.item(i);
|
if (MARKERS.contains(node.getLocalName()) && ("http://schemas.openxmlformats.org/wordprocessingml/2006/main".equals(node.getNamespaceURI())
|
|| "http://purl.oclc.org/ooxml/wordprocessingml/main".equals(node.getNamespaceURI()))) {
|
node.getParentNode().removeChild(node); changed = true;
|
}
|
}
|
return changed;
|
}
|
private static byte[] serialize(Document doc) throws Exception {
|
TransformerFactory factory = TransformerFactory.newDefaultInstance();
|
factory.setFeature(XMLConstants.FEATURE_SECURE_PROCESSING, true);
|
factory.setAttribute(XMLConstants.ACCESS_EXTERNAL_DTD, ""); factory.setAttribute(XMLConstants.ACCESS_EXTERNAL_STYLESHEET, "");
|
ByteArrayOutputStream output = new ByteArrayOutputStream();
|
factory.newTransformer().transform(new DOMSource(doc), new StreamResult(output)); return output.toByteArray();
|
}
|
private record CleanedFile(String name, String originalFilename, String contentType, byte[] bytes) implements MultipartFile {
|
public String getName() { return name; }
|
public String getOriginalFilename() { return originalFilename; }
|
public String getContentType() { return contentType; }
|
public boolean isEmpty() { return bytes.length == 0; }
|
public long getSize() { return bytes.length; }
|
public byte[] getBytes() { return bytes.clone(); }
|
public InputStream getInputStream() { return new ByteArrayInputStream(bytes); }
|
public void transferTo(File dest) throws IOException { Files.write(dest.toPath(), bytes); }
|
}
|
}
|
