0.13.1

This bugfix release renews the support for the version 9 of RNACentral, adding integrity checks to it. This let us simplify the transfers, as we can assume from now on that the files from all supported versions contain metadata with hash information.

Assets 6

12 Nov 11:50

agarciamontoro

v0.13.0

ba4dfb9

db.rnacentral v0.13.0

0.13.0

This version refactors the whole data release process:

aws-scala-tools was removed and s3 and files were added as new dependencies.
The release process is now local: no EC2 instances are run when releasing a new version of the package.
The release process was refactored out of tests and reimplemented in an object on the main code.
The definition of the data was slightly improved.
The old data versions were removed from the code and the current test versions are being uploaded to an unstable subdirectory in S3.
A new test was implemented to check that all supported versions are in S3 as expected.

From now on, the data release process should look something like the following:

Add a new object that extends Version in src/main/scala/data.scala; e.g., case object _11_0 extends Version("11.0")
Perform the mirror through the method release.mirrorNewVersion; e.g., through an sbt console:

scala> import ohnosequences.db.rnacentral._
scala> release.mirrorNewVersion(Version._10_0, new java.io.File("~/localDir"))

Prepare a new release of the repository by writing release notes in the corresponding file in notes; e.g. in notes/0.14.0.markdown
Make the release of the repository, that checks that all declared versions of the data exist and are correct. This is done through the sbt release task; e.g., with sbt release minor

Assets 6

30 Aug 10:41

agarciamontoro

v0.12.2

503d15d

db.rnacentral v0.12.2

0.12.2

Up to RNACentral v7, the sequences and ID mappings files were aligned. From v8, the ID mappings file is no longer ordered.

This release fixes the issues caused by this change, modifying the code to take this into account and removing the tests that checked that the files were actually aligned.

However, this fix has increased both the time and the memory needed to run the ReleaseOnly tests, as the whole ID mappings is now stored in-memory as a map linking the RNA IDs with the annotations.

For more information, see pull request #46.

Assets 6