[libcamera-devel] [PATCH] py: Improve print methods for Transform and ColorSpace objects

[David Plowman]

Hi Laurent

Thanks for the comments.

On Wed, 18 May 2022 at 11:02, Laurent Pinchart
<laurent.pinchart at ideasonboard.com> wrote:
>
> Hi David,
>
> Thank you for the patch.
>
> On Wed, May 18, 2022 at 10:08:01AM +0100, David Plowman via libcamera-devel wrote:
> > They should now print out their meaningful string representations
> > instead of "object at <address>".
> >
> > Signed-off-by: David Plowman <david.plowman at raspberrypi.com>
> > ---
> >  src/py/libcamera/pymain.cpp | 6 ++++++
> >  1 file changed, 6 insertions(+)
> >
> > diff --git a/src/py/libcamera/pymain.cpp b/src/py/libcamera/pymain.cpp
> > index fb89975c..2ba5fcbb 100644
> > --- a/src/py/libcamera/pymain.cpp
> > +++ b/src/py/libcamera/pymain.cpp
> > @@ -567,6 +567,9 @@ PYBIND11_MODULE(_libcamera, m)
> >               .def("__str__", [](Transform &self) {
> >                       return "<libcamera.Transform '" + std::string(transformToString(self)) + "'>";
> >               })
> > +             .def("__repr__", [](Transform &self) {
> > +                     return "<libcamera.Transform '" + std::string(transformToString(self)) + "'>";
> > +             })
>
> Can we do the same as "[PATCH v2 12/13] py: add geometry classes" and
> return a string that would construct an identical Transform object when
> run as Python code ? Same below.

Just to be clear then, you'd prefer a string like (for example)

libcamera.Transform(hflip=False, vflip=False, transpose=False)

?

>
> >               .def_property("hflip",
> >                             [](Transform &self) {
> >                                     return !!(self & Transform::HFlip);
> > @@ -617,6 +620,9 @@ PYBIND11_MODULE(_libcamera, m)
> >               .def("__str__", [](ColorSpace &self) {
> >                       return "<libcamera.ColorSpace '" + self.toString() + "'>";
> >               })
> > +             .def("__repr__", [](ColorSpace &self) {
> > +                     return "<libcamera.ColorSpace '" + self.toString() + "'>";
> > +             })
> >               .def_readwrite("primaries", &ColorSpace::primaries)
> >               .def_readwrite("transferFunction", &ColorSpace::transferFunction)
> >               .def_readwrite("ycbcrEncoding", &ColorSpace::ycbcrEncoding)

And in this case I'd output (for example)

libcamera.ColorSpace(libcamera.ColorSpace.Primaries.Rec709,
libcamera.ColorSpace.TransferFunction.Srgb,
libcamera.ColorSpace.YcbcrEncoding.Rec601,
libcamera.ColorSpace.Range.Full)

?

In this case I could perhaps "optimise" it to

libcamera.ColorSpace.Jpeg()

so I'd check for "short versions" first?

Thanks
David

>
> --
> Regards,
>
> Laurent Pinchart