TheAnachronism/docspell
1
0
Fork 0
mirror of https://github.com/TheAnachronism/docspell.git synced 2025-06-22 10:28:27 +00:00
Code Issues Packages Projects Releases Wiki Activity

Check file integrity

Browse Source
This commit is contained in:
eikek
2022-03-10 22:18:50 +01:00
parent 422c0905dc
commit c1ce0769eb
12 changed files with 285 additions and 19 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
modules/backend/src/main/scala/docspell/backend/BackendApp.scala
View File

@ -92,7 +92,7 @@ object BackendApp {
)
notifyImpl <- ONotification(store, notificationMod)
bookmarksImpl <- OQueryBookmarks(store)
fileRepoImpl <- OFileRepository(queue, joexImpl)
fileRepoImpl <- OFileRepository(store, queue, joexImpl)
} yield new BackendApp[F] {
val pubSub = pubSubT
val login = loginImpl

20
modules/backend/src/main/scala/docspell/backend/JobFactory.scala
View File

@ -15,6 +15,26 @@ import docspell.notification.api.PeriodicQueryArgs
import docspell.store.records.RJob
object JobFactory extends MailAddressCodec {
def integrityCheck[F[_]: Sync](
args: FileIntegrityCheckArgs,
submitter: AccountId = DocspellSystem.account
): F[RJob] =
for {
id <- Ident.randomId[F]
now <- Timestamp.current[F]
job = RJob.newJob(
id,
FileIntegrityCheckArgs.taskName,
submitter.collective,
args,
s"Check integrity of files",
now,
submitter.user,
Priority.High,
Some(FileIntegrityCheckArgs.taskName)
)
} yield job
def fileCopy[F[_]: Sync](
args: FileCopyTaskArgs,
submitter: AccountId = DocspellSystem.account

54
modules/backend/src/main/scala/docspell/backend/ops/OFileRepository.scala
View File

@ -6,27 +6,41 @@
package docspell.backend.ops
import cats.data.OptionT
import cats.effect._
import cats.implicits._
import docspell.backend.JobFactory
import docspell.common.FileCopyTaskArgs
import docspell.backend.ops.OFileRepository.IntegrityResult
import docspell.common._
import docspell.store.Store
import docspell.store.queue.JobQueue
import docspell.store.records.RJob
import scodec.bits.ByteVector
trait OFileRepository[F[_]] {
/** Inserts the job or return None if such a job already is running. */
def cloneFileRepository(args: FileCopyTaskArgs, notifyJoex: Boolean): F[Option[RJob]]
def checkIntegrityAll(part: FileKeyPart, notifyJoex: Boolean): F[Option[RJob]]
def checkIntegrity(key: FileKey, hash: Option[ByteVector]): F[Option[IntegrityResult]]
}
object OFileRepository {
case class IntegrityResult(ok: Boolean, key: FileKey)
def apply[F[_]: Async](
store: Store[F],
queue: JobQueue[F],
joex: OJoex[F]
): Resource[F, OFileRepository[F]] =
Resource.pure(new OFileRepository[F] {
private[this] val logger = docspell.logging.getLogger[F]
def cloneFileRepository(
args: FileCopyTaskArgs,
notifyJoex: Boolean
@ -36,5 +50,43 @@ object OFileRepository {
flag <- queue.insertIfNew(job)
_ <- if (notifyJoex) joex.notifyAllNodes else ().pure[F]
} yield Option.when(flag)(job)
def checkIntegrityAll(part: FileKeyPart, notifyJoex: Boolean): F[Option[RJob]] =
for {
job <- JobFactory.integrityCheck(FileIntegrityCheckArgs(part))
flag <- queue.insertIfNew(job)
_ <- if (notifyJoex) joex.notifyAllNodes else ().pure[F]
} yield Option.when(flag)(job)
def checkIntegrity(
key: FileKey,
hash: Option[ByteVector]
): F[Option[IntegrityResult]] =
(for {
_ <- OptionT.liftF(
logger.debugWith(s"Checking file $key")(_.data("fileKey", key))
)
expectedHash <-
hash.fold(OptionT(store.fileRepo.findMeta(key)).map(_.checksum))(h =>
OptionT.pure[F](h)
)
actualHash <-
OptionT.liftF(
logger.debugWith(s"Calculating new hash for $key")(
_.data("fileKey", key)
) *>
store.fileRepo
.getBytes(key)
.through(fs2.hash.sha256)
.compile
.foldChunks(ByteVector.empty)(_ ++ _.toByteVector)
)
res = IntegrityResult(expectedHash == actualHash, key)
_ <- OptionT.liftF {
if (res.ok) logger.debug(s"File hashes match for $key")
else logger.warnWith(s"File hashes differ for: $key")(_.data("fileKey", key))
}
} yield res).value
})
}