Releases: rsyi/whale
Releases · rsyi/whale
v1.1.2 Cron detection hotfix
Post-release warnings
- Metrics are not super smooth in this version.
Release notes
Nothing major here. One non-blocking bugfix, some doc fixes, code quality improvements, CI/CD improvements.
- When running
wh schedule
(orwh init
) and using a manually installed version of whale (withmake && make install
), the app would give a warning that the crontab is not registered, even if registration successfully occurred (check explicitly withcrontab -e
). This has now been fixed. - Improved panic messages when accessing inaccessible directories.
- Add CONTRIBUTING.md.
- Minor documentation fixes (
alias
suggestion was incorrect). - Added code coverage for rust to the CI/CD pipeline.
v1.1.1 Description hotfix
Post-release warnings
- When running
wh schedule
(orwh init
) and using a manually installed version of whale (withmake && make install
), the app will give a warning that the crontab is not registered, even if registration successfully occurred (check explicitly withcrontab -e
). This is fixed in the next minor release.
Release notes
This fixes #54, enabling Glue (and actually, probably redshift and postgres) to correctly load metadata when descriptions had already been added directly to warehouse tables.
v1.1.0
Post-release warnings
If you have descriptions in glue, this won't work. See v1.1.1 for the bugfix.
Release notes
Not a huge release, but following semantic versioning...
- Added glue support.
- Bugfix: broken
cluster
field in thewh init
onboarding process.
v1.0.0
Post-release warnings
- Glue is not implemented in this version.
- The
wh init
process surfaces acluster/catalog
scraping restriction that actually does nothing. If you are using a warehouse that supports this, use the argumentdatabase
instead ofcluster
.
Original release notes
First official non-alpha/beta release of whale. This release represents a substantial number of changes over metaframe:
- A faster rust backend with
skim
imported as a library (as opposed to using bash with afzf
binary bolted on as a submodule). This represents a move towards full-fledged development and customization of the CLI platform in rust. We've already rewritten the CLI commands in rust, piping into theskim
crate, rather than piping our arguments into a compiled binary, as was the case with metaframe. - A much more sustainable, parseable, and readable (when rendered) markdown format used to save the data. We've move from a two-file system to a one-file backend, reducing number of files, and making it much more intuitive to edit.
- A ~10-20x increase in speed, coming from a combination of (a) caching the table names with each ETL run before loading them into memory (to avoid having to compile a list of all the table stubs with each app run), and (b) reconstructing the backend in rust.
- Deep instructions and commands (
wh git-setup
,wh git-enable
) to make it trivial to use github and github actions (or any other git server + CI/CD service) as the data + scheduling backend. wh schedule
for native support for crontab scheduling: this enables periodic scraping of metadata rather than having to manually runwh pull
orwh etl
.
We've also added support for a number of new [beta] features:
wh run <filename.sql> -w warehouse_name
can be used to run sql queries against the connections you've defined inwh connections
.- Metrics can be defined within the user-generated content of a table stub, and these metrics will be automatically calculated alongside the metadata scraping job.
Finally, we've added some small QOL improvements:
wh init
warehouse connection walkthrough flow, so you don't have to learn the syntax of whale in order to register your connections. Instead, just follow the walkthrough.- Some convenient commands (e.g.
wh config
andwh connections
) to directly access and open files. - Fully functioning brew installation. pyinstaller did not work with bigquery, so we've scrapped this and set up a virtual environment instead (even within homebrew).
- Migration of the repositories to the
dataframehq
namespace.
v1.0.0b1
fix: cd syntax
v0.0.0a31
feat: add snowflake connector. (#26) * feat: add snowflake connector. * chore: set default name to database type, if no name is given.