[
https://issues.apache.org/jira/browse/TIKA-4705?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=18070925#comment-18070925
]
ASF GitHub Bot commented on TIKA-4705:
--------------------------------------
tballison commented on code in PR #2730:
URL: https://github.com/apache/tika/pull/2730#discussion_r3034510123
##########
tika-parsers/tika-parsers-standard/tika-parsers-standard-package/src/test/java/org/apache/tika/parser/RecursiveParserWrapperTest.java:
##########
@@ -174,6 +174,30 @@ public void testTarball() throws Exception {
"/test-documents.tar"), actualEmbeddedPaths);
}
+ @Test
+ public void testNestedTarball() throws Exception {
+ List<Metadata> list = getRecursiveMetadata("test-nested-tarball.tar");
+ List<String> actualInternalPaths =
+ list.stream()
+ .map(m -> m.get(TikaCoreProperties.RESOURCE_NAME_KEY))
+ .collect(Collectors.toList());
+
+ List<String> expectedInternalPaths =
Arrays.asList("test-nested-tarball.tar",
+ "folderWithinTgz/testTXT.txt",
+ "nested.tar",
+ "folderContainingTgz/inner/nested.tgz");
+ assertEquals(expectedInternalPaths, actualInternalPaths);
+
+ List<String> actualEmbeddedPaths =
+ list.stream()
+ .map(m -> m.get(TikaCoreProperties.EMBEDDED_RESOURCE_PATH))
+ .collect(Collectors.toList());
+ assertEquals(Arrays.asList(null,
+ "/nested.tgz/nested.tar/testTXT.txt",
+ "/nested.tgz/nested.tar",
+ "/nested.tgz"), actualEmbeddedPaths);
+ }
Review Comment:
Add test for `internalPaths`?
There are three things we care about with this change: resource names
(should be just the file name), embedded resource path and the internal path.
> resourceName of tar file in nested tarball should not contain tarball's
> parent directories
> ------------------------------------------------------------------------------------------
>
> Key: TIKA-4705
> URL: https://issues.apache.org/jira/browse/TIKA-4705
> Project: Tika
> Issue Type: Improvement
> Reporter: Iachimoe
> Priority: Major
>
> Example structure:
> test-nested-tarball.tar contains:
> folderContainingTgz/inner/nested.tgz
>
> The resource name for nested.tgz would be
> `folderContainingTgz/inner/nested.tgz` , which is consistent with the general
> behaviour for nested archives (e.g. zips).
> However, if nested.tgz does not contain metadata specifying the name of the
> nested file within, then that file will have a resourceName of
> `folderContainingTgz/inner/nested.tar`. This is inconsistent with how other
> nested archives behave, because parent folders should are generally only
> included if they relate to the immediate parent archive. The parent archive
> of nested.tgz in this example is test-nested-tarball.tar , and that is why it
> makes sense for the folders to be included. However, the parent archive of
> nested.tar is nested.tgz , and there is no folder called folderContainingTgz
> within nested.tgz .
>
> Draft pull request with a unit test that hopefully makes the issue clear, and
> a proposed fix at https://github.com/apache/tika/pull/2730/changes
>
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
