Skip to content

CHANGELOG.md

Latest commit

 

History

History
271 lines (218 loc) · 8.83 KB

CHANGELOG.md

File metadata and controls

271 lines (218 loc) · 8.83 KB

Changelog

1.1

1.1.0

  • std feature flag: Enable to use std instead of libm as the math backend.
  • Reduce package size.

1.0

1.0.5

  • Fix no_std: Update num-lazy to v0.4.1 and num-traits to v0.2.19.
  • Refactor the example in README.
  • Add no-std category for crates.io.

1.0.4

  • Update dependencies.

1.0.3

  • Add DOI link to the published paper in JOSS.

1.0.2

  • Ellip is no_std by default.
  • Deprecate no_std feature flag.

1.0.1

Bug Fixes

  • Fix integer overflow in ellipk and ellipe.

1.0.0

  • First stable release of Ellip.

0.5

0.5.9

Minor Fixes

  • Fix the tables in the Testing section of README.

0.5.7

Improvements

  • Add reproducibility section in README.
  • Improve testing documentation.

Tests

  • Add test on macOS and Windows in the CI.
  • Fix the elliprj test for macOS and Windows.

Others

  • Use the 'workspace' project structure.
  • All the children packages now follow the version of ellip.
  • Use cargo-release for the workflow.

0.5.6

Tests

  • Update the error definition, combining symmetric relative error and absolute error.
  • Modify Wolfram test data for elliprj, preventing close symmetric parameter values in the PV domain.

0.5.5

Minor Fixes

  • Fix 3D plots URLs.

0.5.4

Improvements

  • Serve interactive plots through github.io.
  • Automatically generate 3D plots.

Others

  • Upgrade plotly to 0.13.5
  • Use plotly_static instead of kaleido as it is deprecated.

0.5.3

Improvements

  • Improve accuracy of jacobi_zeta and heuman_lambda by using cos for numerical stability.
  • Reduce crate size.

Tests

  • Use exact value of machine epsilon in Wolfram scripts.
  • Fix the percentile calculation for the error report.

0.5.2

Improvements

  • heuman_lambda: Improve performance.
  • Restructure jacobi_zeta and heuman_lambda into misc module.

0.5.1

Bug Fixes

  • Fix incorrect heuman_lambda implementation.

Improvements

  • Reduce crate size.

0.5.0

New Features

  • no_std flag: Support no_std environment.
  • Parallelization available on ellip-rayon crate.

Bug Fixes

  • Fix graphs cutting off at the border in the documentation.

0.4

0.4.0

Breaking Changes

  • BulirschConst: The trait signature changed to BulirschConst<T>.
  • bulirsch functions are bounded by T: Float + BulirschConst<T> trait.

Bug Fixes

  • ellippiinc: Fix segfault when m sin²(phi) > 1. Early catch and return error instead.
  • Fix error message when m > 1 for ellipk, ellipe, ellippi, and ellipd.
  • Fix graphs cutting off at the border in the documentation.

New Features

  • jacobi_zeta and jacobi_zeta_unchecked: Jacobi Zeta function (Z)
  • heuman_lambda and heuman_lambda_unchecked: Heuman Lambda function (Λ0).
  • ellippiinc_bulirsch: Faster implementation of ellippiinc.
  • unstable flag: Access features that might be changed in the future.
  • Add unchecked version for carlson functions, bulirsch functions, ellipeinc and ellippi under the unstable flag.
  • *_with_const: Control the precision of bulirsch functions using BulirschConst trait.
  • DefaultPrecision and HalfPrecision struct for BulirschConst trait available under unstable flag.

Tests

  • Add Wolfram test data for f32 datatype.
  • Add test data for jacobi_zeta and heuman_lambda.

0.3

0.3.7

Bug Fixes

  • ellippiinc: Fix negative m domain where n < 0 and m <= n.

Improvements

  • Improve performance across all functions by removing unnecessary checks.
  • Improve and fix error messages.
  • Add test summary and benchmark data in README.
  • el3: Improve accuracy.
  • elliprj: Slightly increase accuracy of the p.v. cases.

Tests

  • Add Wolfram test data for ellipd and ellipdinc.

0.3.6

Bug Fixes

  • ellipdinc: Fix incorrect result when m = -inf.
  • el3: Fix NAN when p is small and kc = 0.

Improvements

  • Improve performance of legendre functions by lazy checking and el3 by using array instead of Vec.
  • Handle special cases and edge cases (infinity, nan, etc.) for all carlson functions.
  • Increase test coverage.

0.3.5

Deprecations

  • Deprecate BulirschConst: The trait signature will be changed from BulirschConst to BulirschConst<T> and the trait will be relocated under unstable flag.

Bug Fixes

  • cel1: Fix incorrect values in negative kc cases.
  • el3: Fix incorrect values in pm > 0.5 branch.

Improvements

  • Loosen the bounds of the bulirsch functions.
  • Improve the performance of the bulirsch functions by lazy checking.
  • Handle special cases and edge cases (infinity, nan, etc.) for all bulirsch functions.
  • Add special cases documentation for bulirsch functions.
  • Improve and fix documentation.
  • Reduce package size.

Tests

  • Increase test coverage.
  • Include bulirsch functions in the benchmark module.
  • Use native Wolfram functions to generate test data for cel1, el1, and el3.
  • Add error report for cel1 and cel2.

0.3.4

Bug Fixes

  • ellippiinc: Fix not returning error when n sin²φ = 1.

Improvements

  • Handle special cases and edge cases (infinity, nan, etc.) for all legendre functions.
  • Add special cases documentation for legendre functions.
  • Increase test coverage.
  • Report code coverage.
  • Move examples into separate crates, reducing dev dependencies and test time.

0.3.3

Changes

  • Use numeric_literals to parse floats within the functions to make the code more readable.

0.3.2

Improvements

  • Reduce package size.
  • Use custom error type StrErr instead of &'static str to improve readability.
  • Format documentation.

0.3.1

Improvements

  • Improve documentation and add graphs and examples for each function.
  • Improve the precision of the ellipk, ellipf, ellipe, ellipeinc, ellippiinc, and elliprj functions.
  • Improve the speed of ellippiinc function.
  • el3: Add error when parameter goes out of the function's range (|kc| > 0 for p < 0) and improved the error message when 1 + px² equals zero.

New Features

  • Add assert_close function in util module.

Bug Fixes

  • ellipd: Fix incorrect answers for m < 0. The function also returns infinity instead of throwing error at m = 1.
  • Fix the domains for ellipf and ellipeinc.
  • el3: Fix index out of bound in small el3 cases.
  • ellippiinc: Fix infinite loop leading to stack overflow.

0.3.0

New Features

  • ellippi: Complete elliptic integral of the third kind.
  • ellippiinc: Incomplete elliptic integral of the third kind.
  • ellipd: Complete elliptic integral of Legendre's type.
  • ellipdinc: Incomplete elliptic integral of Legendre's type.
  • cel1: Complete elliptic integral of the first kind in Bulirsch's form.
  • cel2: Complete elliptic integral of the second kind in Bulirsch's form.
  • el1: Incomplete elliptic integral of the first kind in Bulirsch's form.
  • el2: Incomplete elliptic integral of the second kind in Bulirsch's form.
  • el3: Incomplete elliptic integral of the third kind in Bulirsch's form.
  • BulirschConst trait controls the precision of cel and el functions.
  • All functions now support generic Float.

Bug Fixes

  • elliprf: Fix incorrect result for RF(x,x,0).
  • elliprj and elliprc: Fix losing precision due to ln(1+x).
  • ellipe: Fix incorrect result during internal condition |t| > 10.

Changes

  • Add dependency num-traits and num-lazy.

0.2

0.2.1

Improvements

  • Reduce crate size by removing the logo and test data from the source. To perform full test, download the test data from the repository.

0.2.0

New Features

  • cel: Bulirsch's general complete elliptic integral.
  • elliprg: Symmetric elliptic integral of the second kind.
  • elliprj: Symmetric elliptic integral of the third kind.

Improvements

  • elliprf: Add special cases handling to improve performance and accuracy.
  • elliprd: Use Boost Math implementation to improve accuracy.
  • ellipe: Explicitly make internal functions inline.

Changes

  • Change error messages in ellipk, ellipf, ellipe, and ellipeinc, to describe without mathematical notations.
  • Split error message in elliprd.

Bug Fixes

  • Fix domain error message in elliprd.

Others

  • Add Ellip logo
  • Add links to references in the documentations.

0.1

0.1.2

Bug Fixes

  • Fix ellipe logic in negative m cases.

0.1.1

Others

  • Update README to reflect the crate's functionalities.
  • Add CHANGELOG.

0.1.0

New Features

  • ellipf: Incomplete elliptic integral of the first kind.
  • ellipeinc: Incomplete elliptic integral of the second kind.
  • ellipk: Complete elliptic integral of the first kind.
  • ellipe: Complete elliptic integral of the second kind.
  • elliprf: Symmetric elliptic integral of the first kind.
  • elliprd: Degenerate elliptic integral of the third kind.
  • elliprc: Degenerate elliptic integral of RF.