sepinf-inc#2286: link ufed message to correspondent chat part

aberenguel · Aug 15, 2024 · dcd7e49 · dcd7e49
1 parent 04b3cfc
commit dcd7e49
Show file tree

Hide file tree

Showing 2 changed files with 14 additions and 10 deletions.
diff --git a/iped-app/resources/config/conf/ParserConfig.xml b/iped-app/resources/config/conf/ParserConfig.xml
@@ -291,11 +291,6 @@
                 <param name="extractEntries" type="bool">false</param>
             </params>
         </parser>
-        <parser class="iped.parsers.ufed.UFEDChatParser">
-            <params>
-				<param name="extractMessages" type="bool">true</param>
-            </params>
-        </parser>
         <parser class="iped.parsers.telegram.TelegramParser">
             <params>
 				<param name="extractMessages" type="bool">true</param>
@@ -328,6 +323,7 @@
         <parser class="iped.parsers.ufed.UFEDChatParser">
             <!--<mime-exclude>application/x-ufed-chat-whatsapp</mime-exclude>-->
             <params>
+            	<param name="extractMessages" type="bool">true</param>
                 <param name="minChatSplitSize" type="int">6000000</param>
             </params>
         </parser>

diff --git a/iped-parsers/iped-parsers-impl/src/main/java/iped/parsers/ufed/UFEDChatParser.java b/iped-parsers/iped-parsers-impl/src/main/java/iped/parsers/ufed/UFEDChatParser.java
@@ -241,11 +241,11 @@ public void parse(InputStream inputStream, ContentHandler handler, Metadata meta
                     ByteArrayInputStream chatStream = new ByteArrayInputStream(bytes);
                     extractor.parseEmbedded(chatStream, handler, chatMetadata, false);
                     bytes = nextBytes;
-                }
-            }
 
-            if (extractMessages) {
-                extractMessages(itemMgs, virtualId, handler, extractor);
+                    if (extractMessages) {
+                        extractMessages(itemMgs, subList, virtualId, handler, extractor);
+                    }
+                }
             }
 
         } catch (Exception e) {
@@ -258,9 +258,17 @@ public void parse(InputStream inputStream, ContentHandler handler, Metadata meta
 
     }
 
-    private void extractMessages(List<IItemReader> messages, String virtualId, ContentHandler handler,
+    private void extractMessages(List<IItemReader> messages, List<UfedMessage> subList, String virtualId, ContentHandler handler,
             EmbeddedDocumentExtractor extractor) throws SAXException, IOException {
+
+        HashSet<Long> idsInSubList = subList.stream().mapToLong(UfedMessage::getId).boxed().collect(Collectors.toCollection(HashSet::new));
+
         for (IItemReader msg : messages) {
+
+            if (!idsInSubList.contains((long) msg.getId())) {
+                continue;
+            }
+
             Metadata meta = msg.getMetadata();
             meta.set(TikaCoreProperties.TITLE, msg.getName()); // $NON-NLS-1$
             meta.set(BasicProps.ID, Integer.toString(msg.getId()));