Minimal, idiomatic, stream-based Scala interface for key/value store implementations. It provides abstractions for S3-like key/value store backed by different persistence mechanisms (i.e. S3, FileSystem, sftp, etc).
fs2-blobstore is deployed to maven central, add to build.sbt:
libraryDependencies ++= Seq(
"com.lendup.fs2-blobstore" %% "core" % "0.6.+",
"com.lendup.fs2-blobstore" %% "sftp" % "0.6.+",
"com.lendup.fs2-blobstore" %% "s3" % "0.6.+"
)
core
module has minimal dependencies and only provides FileStore
implementation.
sftp
module provides SftpStore
and depends on Jsch client.
s3
module provides S3Store
and depends on AWS S3 SDK
The goal of the Store interface is to
have a common representation of key/value functionality (get, put, list, etc) as
streams that can be composed, transformed and piped just like any other fs2.Stream
or fs2.Sink
regardless of the underlying storage mechanism.
This is especially useful for unit testing if you are building a S3 or SFTP backed system as you can provide a filesystem based implementation for tests that is guaranteed to work the same way as you production environment.
The three main activities in a key/value store are modeled like:
def list(path: Path): fs2.Stream[F, Path]
def get(path: Path, chunkSize: Int): fs2.Stream[F, Byte]
def put(path: Path, contentLength: Long): fs2.Sink[F, Byte]
Note that list
and get
are modeled as streams since they are reading
(potentially) very large amounts of data from storage, while put
is
represented as a sink of byte so that any stream of bytes can by piped
into it to upload data to storage.
import blobstore.implicits._
StoreOps and
PathOps provide functionality on
both Store
and Path
for commonly performed tasks (i.e. upload/download a
file from/to local filesystem, collect all returned paths when listing, composing
paths or extracting filename of the path).
Most of these common tasks encapsulate stream manipulation and provide a simpler interface that return the corresponding effect monad. These are also very good examples of how to use blobstore streams and sink in different scenarios.
All store implementations must support and pass the suite of tests in
AbstractStoreTest.
It is expected that each store implementation (like s3, sftp, file) should
contain the Store
implementation and at least one test suite that inherits
from AbstractStoreTest
and overrides store and root attributes:
class MyStoreImplTest extends blobstore.AbstractStoreTest {
override val store: blobstore.Store[cats.effect.IO] = MyStoreImpl( ... )
override val root: String = "my_store_impl_tests"
}
This test suite will guarantee that basic operations are supported properly and
consistent with all other Store
implementations.
Running Tests:
Tests are set up to run via docker-compose:
docker-compose run --rm sbt "testOnly * -- -l blobstore.IntegrationTest"
This will start a minio (Amazon S3 compatible
object storage server) and SFTP containers and run all tests not annotated as
@IntegrationTest
.
Yes, we understand SftpStoreTest
and S3StoreTest
are also integration tests
because they connect to external services, but we don't mark them as such because
we found these containers that allow to run them along unit tests and we want to
exercise as much of the store code as possible.
Currently, tests for SftpStore
and BoxStore
are annotated with @IntegrationTest
because: (1) SFTP tests fail to run against sftp container in travis, and (2) we
have not found a box docker image. To run BoxStore
integration tests locally
you need to provide env vars for BOX_TEST_BOX_DEV_TOKEN
and BOX_TEST_ROOT_FOLDER_ID
.
Run box/sftp tests with:
sbt box/test
docker-compose run --rm sbt sftp/test
Note: this will exercise AbstractStoreTest
tests against your box.com account.
blobstore.Path
is the representation of key
in the key/value store. The key
representation is based on S3 that has a root
(or bucket) and a key
string.
When functions in the Store
interface that receive a Path
should assume that only
root and key values are set, there is no guarantee that the other attributes of Path
would be filled: size, isDir, lastModified. On the other hand, when a Store
implements
the list function, it is expected that all 3 fields will be present in the response.
By importing implicit PathOps
into the scope you can make use of path composition /
and filename
function that returns the substring of the path's key after the last path
separator.
NOTE: a good improvement to the path abstraction would be to handle OS specific separators when referring to filesystem paths.
import blobstore.Store, blobstore.fs.FileStore
import java.nio.file.Paths
import cats.effect.IO
val store: Store[IO] = FileStore[IO](Paths.get("tmp/"))
AmazonS3
client:
import blobstore.Store, blobstore.s3.S3Store
import com.amazonaws.services.s3.transfer.TransferManagerBuilder
import cats.effect.IO
val store: Store[IO] = S3Store[IO](TransferManagerBuilder.standard().build())
SftpStore backed by
SFTP server with Jsch client. It requires a
connected ChannelSftp
:
import blobstore.Store, blobstore.sftp.SftpStore
import com.jcraft.jsch.{ChannelSftp, JSch}
import cats.effect.IO
val jsch = new JSch()
val session = jsch.getSession("sftp.domain.com")
session.connect()
val channel = session.openChannel("sftp").asInstanceOf[ChannelSftp]
channel.connect(5000)
val store: Store[IO] = SftpStore("root/server/path", channel)
BoxStore backed by a BoxAPIConnection, which has multiple options for authentication. This requires that you have a Box app set up already. See Box SDK documentation for more details:
import blobstore.Store, blobstore.box.BoxStore
import com.box.sdk.BoxAPIConnection
val api = new BoxAPIConnection("myDeveloperToken")
val store: Store[IO] = BoxStore[IO](api, "rootFolderId")