Mir vs. Numpy: Reworked!
Andre Pany
andre at s-e-a-p.de
Sat Dec 5 13:17:47 UTC 2020
On Saturday, 5 December 2020 at 07:04:59 UTC, 9il wrote:
> On Thursday, 3 December 2020 at 16:50:39 UTC, Andre Pany wrote:
>> On Thursday, 3 December 2020 at 16:27:59 UTC, 9il wrote:
>>> [...]
>>
>> Hi Ilya,
>>
>> Thanks a lot for sharing the update. I am currently working on
>> porting a python package called FMPY to D. This package makes
>> usage of numpy and I hope I can use MIR here.
>
> Probably you may want to express FMI entities as Algebraic
> types rather than classes. mir.algebraic can be really helpful
> here
>
> http://mir-core.libmir.org/mir_algebraic.html
>
>> Somehow it is hard to get started to learn MIR. What maybe
>> could help python developers is to have some articles showing
>> numpy coding and side by side the equivalent MIR coding.
>
> It is hard for me to write articles. I will try to write a
> small one this year, but it would be Mir only. Maybe this
> benchmark can be used as an example and if one wishes to write
> a side-by-side comparison with NumPy I would be happy to
> comment and explain the D implementation and what it is doing
> internally.
>
>> What I miss in MIR is a function to read and write CSV files.
>> Is s.th. like numpy.genfromtxt planned?
>
> Unlikely I would add it but can do a code review.
>
> Currently, we can load/safe NumPy binary data with numir
>
> https://libmir.github.io/numir/io.html
>
> Kind regards,
> Ilya
Thanks a lot for the tipps. I will work through the documentation
of mir_algebraic. Currently I follow the strategy to have the D
coding as similiar as possible to the python coding. Fmpy is in
active development and I want to backport changes easily.
I totally missed the fact that MIR can load / safe numpy binary
data. This is really great.
Kind regards
Andre
More information about the Digitalmars-d-announce
mailing list