Release note for v0.3.0

[heronshoes]

RedAmber v0.3.0 has released 🎉

Ready for Ruby 3.2 and dropped Ruby 2.7 .

This release is ready for Ruby 3.2 . I added CI test from 3.2.0-preview3 and no issue has found.
I dropped support for Ruby 2.7 in this release. The reasons are;

• Ruby 3 has a lot of performance improvements.
• EOL of 2.7 will coming soon. There are no positive reason to use 2.7 especially for performance oriented data processing.
• 2.7 shows warning for pattern matching used frequently in RedAmber.
• 2.7 can't deal Hash and keyword arguments in join.
  • API of join is #join(other, join_key = nil, type: :inner, suffix: '.1'). While join_key may be a Hash of { left: :left_key, right: :right_key }. Also join_key may be nil (implicit common key is used).
    I can't run 2.7 as same manner as over 3.0 in this case.

So I decided to drop 2.7 .

Big clean-up (大掃除) of code and fastest RedAmber ever

RedAmber 0.2.3 was the most feature-rich version and therefore the slowest.

I completely overhauled the code using the performance measurements introduced in 0.2.3 and was able to achieve an 11.7x performance improvement over v0.2.0 and a 12.2x improvement over v0.2.3.

Also, I refined the code and added missing tests. Test coverage was 99.1% in v0.2.3, I pulled up to 100% in v0.3.0 . I will maintain test coverage.

Improvement of CI

This version includes installation instructions and CI tests for fedra by Benson Muite and YARD documentation generation workflow by kojix2. I appreciate their contributions.

Stats from 0.2.3 to 0.3.0

• There was 2803 insertions and 800 deletions.
• 519 tests, 1424 assertions => 605 tests, 1651 assertions

Changelog of this release is below. Your help is welcome!

CHANGELOG [0.3.0] - 2022-12-18

• Breaking change

  • Supported Ruby version has changed from 2.7 to 3.0
    • Upgrade minimum supported/required version of Ruby from 2.7 to 3.0 (Improve DataFrame#join #159, Prepare 0.3.0 #160)

• Bug fixes

  • Add check with #key? in DataFrame#method_missing (Refactor code #140)
  • Delete unnecessary backslash to supress warning in unary functions (Refactor code #140)
  • Fix syntax in code_climate.yml (Send test coverage report to Code Climate #144)
  • Temporary disable simplecov test report (Temporary disable simplecov test report #149)
  • Change Vector#[] to return Array or scalar (Update Class Vector #148)
  • Add missing simplecov HTML formatter (Update Class Vector #148)
  • Change return value of DataFrame#save to self (Prepare 0.3.0 #160)
    • Originally reported by kojix2.

• New features and improvements

  • Update Vector#take to accept block (Update Class Vector #148)
  • Add properties of list Vectors (Update Class Vector #148)
  • Add Vector#split, #split_to_column, #split_to_row (Update Class Vector #148)
  • Add Vector#merge (Update Class Vector #148)

• Refactoring

  • Refactor code (Refactor code #140)
    • Add DataFrame.create as a faster constructor
    • Refactor DataFrame.new using refinements and duck typing
    • Refactor Vector.new using refinements and duck typing
    • Add Vector.create as a faster constructor
    • Refactor Group
    • Refactor DataFrame#pick/#drop by refining Array
    • Refactor DataFrame#pick/#drop
    • Refactor nil treatment in pick/drop
    • Refactor DataFrame#pick/#drop using new parser
    • Refactor DataFrame#[]
    • Refactor Vector#[], #take, #filter by updating parser
    • Add for_keys option to parse_args
    • Refactor Vector properties by refinements for Arrow::Array
    • Refactor DataFrame selectable using Arrow::Array refinements instead of Vector methods
    • Refactor DataFrame#assign
  • Refine error message in DataFrame#to_long/to_wide Refine error message in DataFrame#to_long/to_wide #143)
  • Refactor Vector#take/filter returns arrow array (Update Class Vector #148)
  • Change LineLength in cop from 120 to 90 (Refactor line length #152)
  • Refine DataFrame combinable (join) operations (Improve DataFrame#join #159)
    • Refine DataFrame#join effectively using outputs options
    • Simplify DataFrame set operations

• Improve in tests/CI

  • Tests

    • Update benchmark using 0.2.3 (Bump version of HEAD #138)
    • Update benchmark basic#02/pick by [] (Refactor code #140)
    • Update benchmark contexts and loop_count (Refactor code #140)
    • Add benchmark for vector (Refactor code #140)
    • Add tests for refinements (Refactor code #140)
    • Add benchmark for the series of DataFrame operations (Refactor code #140)
    • Add missing test for tdr and dictionary (Refactor code #140)
    • Add missing test for group#method with foreign key (Refactor line length #152)
    • Add missing test for set operations and natural join (Refactor line length #152)
    • Add missing test for DataFrame#[] with selecting by Array of illegal type' (Refactor line length #152)
    • Add missing test for DataFrame#assign when assigner size is mismatch (Refactor line length #152)
    • Accept Hash as join keys in DataFrame join methods (Improve DataFrame#join #159)

  • Cops

    • Refactor/clean rubocop.yml (Bump version of HEAD #138)

  • CI

    • Support Ruby 3.2 in CI test (Support Ruby 3.2 in CI test #141)

    • Send test coverage report to Code Climate (Send test coverage report to Code Climate #144)

    • Add workflow test on Fedora (Test on Fedora #151)

      • Thanks to Benson Muite.

    • Add workflow to generate document (Add workflow to generate document #153)

      • Thanks to kojix2.

    • Support Code Climate test coverage report in CI (Revert Code Climate report in CI #155)

• Documentation

  • Add YARD in data_frame.rb (Refactor code #140)

  • Fix YARD document in the code (Refactor code #140)

  • Add Code Climate badges of maintainability and coverage (Send test coverage report to Code Climate #144)

  • Add installation for Fedora in README (Add installation for Fedora, Workaround for #146 #147)

    • Thanks to Benson Muite.

  • Add Vector#split/merge in Vector.md (Update Class Vector #148)

  • Fix codeclimate badges in README (Revert Code Climate report in CI #155)

  • Update YARD in DataFrame join methods (Improve DataFrame#join #159)

  • Update jupyter notebook '89 examples of Redamber' (Prepare 0.3.0 #160)

• Thanks

  • Benson Muite
  • kojix2