Kaskada 0.6.0-a.4

✨ Features & Enhancements

• initial logical to physical compile (#792)
• support lists in spread zip (#793)
• add select transform (#789)
• support time unit via preparing parquet from python (#786)
• Expose DFG and query plans on Timestream (#780)
• allow periodic filtering of aggregations using a tumbling window (#758)
• add parquet source to python ffi (#749)
• Async back-pressure on adding to a source (#746)

🐞 Bug fixes

• temporarily ignore parquet source tests to build windows (#796)
• uses URL instead of string to construct path to file (#797)
• make file paths platform-independent (#795)
• drains output channel when batching results (#763)
• fix shift to batch bounds (#770)

🛠️ Other improvements

20 changes

• move prepared files to ~/.cache dir (#781)
• import the community page (#791)
• Update Arrow version (#784)
• Simplify expressions in physical plan (#782)
• use new prepare code and remove some old (#778)
• added new jsonl-file source (#773)
• fixed issue when subsort is provided (#771)
• fix typo in docs (#769)
• added substring, json, len, is_valid methods (#767)
• locked python client to engine@v0.11.0 (#764)
• Add an example reading from Reddit (#757)
• Introduce logical plan creation (#756)
• Flesh out scripts for the SQL benchmark (#762)
• remove print in parquet source method (#760)
• replace first/last with generic aggregators (#755)
• Doc fixes to align to new API and release versions (#750)
• Update pyproject.toml (#753)
• Add an example showing how to consume from BlueSky (#751)
• tweaks to bluesky example (#752)
• remove unused file (#742)

Thank you to all our contributors for making this release possible!
@bjchambers, @epinzur, @jordanrfrazier and @kerinin