file: unzip introduced #2692
file: unzip introduced #2692
Conversation
|
Nice! Some docs would be required before merging. |
|
I added docs, but can't test it, for some reason the google protobuff things are not compile on my machine. |
| import java.io.{File, FileInputStream} | ||
| import java.util.zip.{ZipEntry, ZipInputStream} | ||
|
|
||
| case class ZipArchiveMetadata(name: String) |
There was a problem hiding this comment.
This class is user-facing API and should preferably live in akka.stream.alpakka.file instead. Add a getName for Java folks.
| @@ -106,6 +107,23 @@ public void flowShouldCreateZIPArchive() throws Exception { | |||
| Map<String, ByteString> unzip = archiveHelper.unzip(resultFileContent); | |||
|
|
|||
| assertThat(inputFiles, is(unzip)); | |||
| Path target = Files.createTempDirectory("alpakka-tar-"); | |||
There was a problem hiding this comment.
| Path target = Files.createTempDirectory("alpakka-tar-"); | |
| Path target = Files.createTempDirectory("alpakka-zip-"); |
| } | ||
| } | ||
|
|
||
| @InternalApi class ZipSource(f: File, chunkSize: Int) |
There was a problem hiding this comment.
@tg44 I see that ZipSource "only" handles Files. However what about this scenario: In a Play app a user uploads a zip file and I want to directly, without storing it on disk, unpack it and stream its content(s) to e.g. S3 buckets. This could probably be done in a Play Body parser. I guess ZipSource would need to handle Source to achieve that?
What do you think? Would it be somehow possible to adapt your code to support such a scenario?
Thanks!
There was a problem hiding this comment.
The current source works with file seek. I needed to parallel access files. Without File and seeking, you need to consume the "subSources", and you can't use parallelism, which is in my opinion makes the source hard to use, and easy to misuse. (I think the tar flow works like this right now.)
There was a problem hiding this comment.
I totally understand your approach. I had a look at the tar reading example now, and I think it would be good if we could have the same for zip reading as well (in addition to your implementation). What also bothers me a bit is that for writing zip files as well as for read/writing tar file we work with sources and sinks, but for reading zip files we only allow Files now.
There was a problem hiding this comment.
Ofc. we can implement a tar reading like API too, I just haven't needed that :) I think you can easily merge the tar reading and zip file reading logic. (Sorry, but I'm really short on time nowadays, so I can't contribute that in the next weeks.)
docs are missing, the basic test works, ideas welcome :)