Releases: uccross/skyhookdm-arrow
v0.4.1
v0.4.0
SKYHOOK-273: [CI] Add benchmark trigger workflow
v0.3.0
What's new in this release 🎉
-
039b1d7, 9e9d6dd : Rename
RadosParquetFileFormat
toSkyhookFileFormat
to abstract over different real file formats like CSV, IPC, JSON, Parquet. -
07b2f20: Add support for
RadosParquet
to JNI Dataset API bindings and update SkyhookDM deployment scripts to install JARs for the JNI Dataset API. -
2e8598f - Add a thread-safe connection abstraction for
RadosParquet
along with other minor improvements to the APIs. -
64181c0: Rebase SkyhookDM to latest Apache Arrow 5.0.0 release.
-
683fe48: Set
osd_class_dir
to make the CLS library lookup path portable across environments. -
79e3073, a8dd52b, 8805a21, f0285ef, dfd0c79 - Add docstrings to all the new source files added for SkyhookDM.
-
6dc047e - Add Netlify based CD workflow for automatically deploying documentation on every push.
-
b111e7a, e2a5e2a - Add a benchmarking guide for beginners to get started with SkyhookDM benchmarking.
-
b08fe9e, 93c008c : Made Ceph and Skyhook deploy scripts more dynamic and fault tolerant.
Thanks a lot to @YashJipkate , @ljishen , and @Sebastiaan-Alvarez-Rodriguez for their valuable contributions in this release.
v0.2.1-rc1
SKYHOOK-207: [Java] Add optional Java and JNI builds to deploy_skyhoo…
v0.2.0
Changes in v0.2.0
- c0f278b: Upgrade SkyhookDM to Apache Arrow 4.x.
- Send back results in much more faster Lz4 compressed IPC format.
- 25bc170: Flatbuffers to serialize Scan Request.
- Move skyhook-ci workflow to
uccross/skyhookdm-arrow-docker
- 8636266: Fix memory leaks on both the client and CLS sides.
- 39a535c: Modify
SplittedParquetWriter
to be able to split Parquet files by estimating row group sizes based on compression ratio. - Update documentation with architecture document and add utility scripts.
v0.1.1
v0.1.0 - First Release !
Please follow this README for a quick start guide and instructions for installation and setting up the development environment.
Thanks to @jlefevre, @ivotron, @esmaeil-mirvakili, and @kingwind123 for their invaluable contributions.