Merge pull request #203 from guardian/SNYK-JAVA-ORGAPACHECOMMONS-6254296

upgrading tika to 2.9.1 to fix SNYK-JAVA-ORGAPACHECOMMONS-6254296
guardian · Mar 13, 2024 · b2aee07 · b2aee07
2 parents efcd5ae + 9b7a400
commit b2aee07
Show file tree

Hide file tree

Showing 3 changed files with 3 additions and 2 deletions.
diff --git a/backend/app/extraction/archives/ZipExtractor.scala b/backend/app/extraction/archives/ZipExtractor.scala
@@ -32,7 +32,7 @@ class ZipExtractor(scratch: ScratchSpace, ingestionServices: IngestionServices)
     val extractionRootPath = scratch.createWorkingDir(s"zip/${blob.uri.value}/")
     logger.info(s"Running ZIP extractor on '${blob.uri.value}' in temporary working directory '$extractionRootPath'")
 
-    val zipFile = new ZipFile(file)
+    val zipFile = ZipFile.builder().setFile(file).get()
     val builder = IngestionContextBuilder(blob.uri, params)
 
     val result = Either.catchNonFatal {

diff --git a/backend/app/extraction/email/olm/OlmEmailExtractor.scala b/backend/app/extraction/email/olm/OlmEmailExtractor.scala
@@ -31,7 +31,7 @@ class OlmEmailExtractor(scratch: ScratchSpace, ingestion: IngestionServices) ext
 
     val context = IngestionContextBuilder(blob.uri, params)
 
-    val zipFile = new ZipFile(scratchFile)
+    val zipFile = ZipFile.builder().setFile(scratchFile).get()
     logger.info(s"Loaded OLM '${blob.uri.value}'")
 
     try {

diff --git a/build.sbt b/build.sbt
@@ -131,6 +131,7 @@ lazy val backend = (project in file("backend"))
       // (Seems to be OK as of 2.7.0: https://tika.apache.org/2.7.0/parser_guide.html)
       "org.apache.tika" % "tika-parsers-standard-package" % "2.7.0",
       "org.apache.tika" % "tika-core" % "2.7.0",
+      "org.apache.commons" % "commons-compress" % "1.26.0",
       "org.apache.logging.log4j" % "log4j-to-slf4j" % log4jVersion,
       "org.apache.logging.log4j" % "log4j-api" % log4jVersion,
       "org.apache.logging.log4j" % "log4j-core" % log4jVersion,