Towards gh-18867
This issue tracks progress toward the addition of array-API support to scipy.integrate functions. The functions listed below look ready for conversion.
trapezoid ENH: integrate.trapezoid: add array API standard support #21524simpson ENH: integrate.simpson: add array API standard support #21548rombcumulative_trapezoidcumulative_simpsoncubature ENH: integrate.cubature: array API standard support #21473If I were to review, I'd prefer to see one PR per function in the order listed above with only one PR open at a time, learning from each as we go. (If others are willing to review without these conditions, go for it!)
I'd like to work on:
special.logsumexp (needed for qmc_quad, ENH: special.logsumexp: add array API standard support #21149)qmc_quadThese would be pretty substantial undertakings; there would need a pretty dedicated team of author(s)/reviewer(s). I'd suggest holding off on them until some of the lower-hanging fruit is picked. @nmayorov would you be interested in being part of this?
solve_ivpsolve_bvpOf course, adaptive quadrature given a callable rather than points is also a very high priority. We'll see where gh-20252 is at the end of the summer.
I think we could convert these functions, but I'm not sure whether doing so is a high priority. LMK your thoughts.
newton_cotesfixed_quad (IIUC, automatic differentiation would be easier with something non-adaptive like this. I'm assuming that RFC: Multidimensional (cubature) integration #20252 will keep this in mind.)I don't think these will be converted, per se.
quad_vec - either the outcome of RFC: Multidimensional (cubature) integration #20252 or _tanhsinh would be a suitable array-API compatible replacementquad - Fortran, and same as aboveodeint(and related) - Fortrandblquad, tplquad, nquad - only relevant if quad were convertedRetroSearch is an open source project built by @garambo | Open a GitHub Issue
Search and Browse the WWW like it's 1997 | Search results from DuckDuckGo
HTML:
3.2
| Encoding:
UTF-8
| Version:
0.7.4