v0.1.0

Bessels v0.1.0

Closed issues:

• bessely (#2)
• Implement solutions for besselj(nu, x) for integer orders (#3)
• Poor accuracy for middle orders 2-50 for besseli (#9)
• Anything to take from Fortran? (#18)
• Could return complex values for real inputs if argument is negative (#30)

Merged pull requests:

• add Float64 benchmark in readme (#4) (@heltonmc)
• merge Float32 and Float64 implimentations (#5) (@oscardssmith)
• Change besseli0 and besseli1 routines to rational approximation (#6) (@heltonmc)
• Faster implementations of Besselk0 and Besselk1 (#7) (@heltonmc)
• Asymptotic expansion for large orders. Besselk(nu, x) (#8) (@heltonmc)
• Use continued fractions for besseli for medium sized orders 2<nu<100 (#10) (@heltonmc)
• benchmark suite (#11) (@heltonmc)
• update readme, fix typo (#12) (@oscardssmith)
• Better large order approximations for J0,J1, Y0,Y1 (#13) (@heltonmc)
• besselj0 and friends for medium arguments (#14) (@heltonmc)
• faster besselj0 for x<25 (#15) (@oscardssmith)
• Implementation for besselj of any (positive) real order and argument (#20) (@heltonmc)
• optimize U-polynomials and output both besselj and bessely for debye expansion (#23) (@heltonmc)
• Output both bessely and besselj for large arguments (#24) (@heltonmc)
• bessely(nu, x) for positive and negative arguments (#26) (@heltonmc)
• Clean up besselj(nu,x) and add docs (#28) (@heltonmc)
• Besselj and Besselk for all integer and noninteger postive orders (#29) (@heltonmc)
• Update readme (#31) (@heltonmc)
• remove complex output for real inputs (#32) (@heltonmc)