[libcamera-devel] [PATCH v2] py: Add the SensorConfiguration class

[Tomi Valkeinen]

On 20/10/2023 11:17, David Plowman wrote:
> Hi Tomi
> 
> Thanks for that. I had no idea you could write that... the joys of C++!

I think usually with pybind11 you should think in terms of C++. Use the 
C++ types whenever possible, as they're more performant and specific 
than the Python types, and let pybind11 do the conversions between the 
Python and C++ types.

  Tomi

> 
> Would folks like me to send another update?
> 
> David
> 
> On Fri, 20 Oct 2023 at 08:51, Tomi Valkeinen
> <tomi.valkeinen at ideasonboard.com> wrote:
>>
>> On 12/10/2023 11:29, Kieran Bingham via libcamera-devel wrote:
>>> Hi David,
>>>
>>> Quoting David Plowman via libcamera-devel (2023-10-11 15:02:38)
>>>> We provide access to the various fields of the new SensorConfiguration
>>>> class. The class also needs a constructor so that Python applications
>>>> can make one and put it into the CameraConfiguration.
>>>>
>>>> Signed-off-by: David Plowman <david.plowman at raspberrypi.com>
>>>> ---
>>>>    src/py/libcamera/py_main.cpp | 36 ++++++++++++++++++++++++++++++++++++
>>>>    1 file changed, 36 insertions(+)
>>>>
>>>> diff --git a/src/py/libcamera/py_main.cpp b/src/py/libcamera/py_main.cpp
>>>> index 01fb15a9..d2ce3722 100644
>>>> --- a/src/py/libcamera/py_main.cpp
>>>> +++ b/src/py/libcamera/py_main.cpp
>>>> @@ -112,6 +112,7 @@ PYBIND11_MODULE(_libcamera, m)
>>>>
>>>>           auto pyCameraManager = py::class_<PyCameraManager, std::shared_ptr<PyCameraManager>>(m, "CameraManager");
>>>>           auto pyCamera = py::class_<Camera, PyCameraSmartPtr<Camera>>(m, "Camera");
>>>> +       auto pySensorConfiguration = py::class_<SensorConfiguration>(m, "SensorConfiguration");
>>>>           auto pyCameraConfiguration = py::class_<CameraConfiguration>(m, "CameraConfiguration");
>>>>           auto pyCameraConfigurationStatus = py::enum_<CameraConfiguration::Status>(pyCameraConfiguration, "Status");
>>>>           auto pyStreamConfiguration = py::class_<StreamConfiguration>(m, "StreamConfiguration");
>>>> @@ -281,6 +282,40 @@ PYBIND11_MODULE(_libcamera, m)
>>>>                           return ret;
>>>>                   });
>>>>
>>>> +       pySensorConfiguration
>>>> +               .def(py::init<>())
>>>> +               .def_readwrite("bit_depth", &SensorConfiguration::bitDepth)
>>>> +               .def_readwrite("analog_crop", &SensorConfiguration::analogCrop)
>>>> +               .def_property(
>>>> +                       "binning",
>>>> +                       [](SensorConfiguration &self) {
>>>> +                               return py::make_tuple(self.binning.binX, self.binning.binY);
>>>> +                       },
>>>> +                       [](SensorConfiguration &self, py::object value) {
>>>> +                               auto vec = value.cast<std::vector<unsigned int>>();
>>>> +                               if (vec.size() != 2)
>>>> +                                       throw std::runtime_error("binning requires iterable of 2 values");
>>>> +                               self.binning.binX = vec[0];
>>>> +                               self.binning.binY = vec[1];
>>>
>>> This seems reasonable to me. Partially filling in either of binning or
>>> skipping would be invalid.
>>>
>>> Thanks for filling in the rest of the logic:
>>>
>>> Reviewed-by: Kieran Bingham <kieran.bingham at ideasonboard.com>
>>>
>>> Tomi - Are you happy with this?
>>
>> I think it's simpler and better to do:
>>
>> .def_property(
>>          "binning",
>>          [](SensorConfiguration &self) {
>>                  return std::make_tuple(self.binning.binX, self.binning.binY);
>>          },
>>          [](SensorConfiguration &self, std::tuple<uint32_t, uint32_t> value) {
>>                  self.binning.binX = std::get<0>(value);
>>                  self.binning.binY = std::get<1>(value);
>>          })
>>
>>    Tomi
>>
>>
>>>
>>>
>>>
>>>> +                       })
>>>> +               .def_property(
>>>> +                       "skipping",
>>>> +                       [](SensorConfiguration &self) {
>>>> +                               return py::make_tuple(self.skipping.xOddInc, self.skipping.xEvenInc,
>>>> +                                                     self.skipping.yOddInc, self.skipping.yEvenInc);
>>>> +                       },
>>>> +                       [](SensorConfiguration &self, py::object value) {
>>>> +                               auto vec = value.cast<std::vector<unsigned int>>();
>>>> +                               if (vec.size() != 4)
>>>> +                                       throw std::runtime_error("skipping requires iterable of 4 values");
>>>> +                               self.skipping.xOddInc = vec[0];
>>>> +                               self.skipping.xEvenInc = vec[1];
>>>> +                               self.skipping.yOddInc = vec[2];
>>>> +                               self.skipping.yEvenInc = vec[3];
>>>> +                       })
>>>> +               .def_readwrite("output_size", &SensorConfiguration::outputSize)
>>>> +               .def("is_valid", &SensorConfiguration::isValid);
>>>> +
>>>>           pyCameraConfiguration
>>>>                   .def("__iter__", [](CameraConfiguration &self) {
>>>>                           return py::make_iterator<py::return_value_policy::reference_internal>(self);
>>>> @@ -293,6 +328,7 @@ PYBIND11_MODULE(_libcamera, m)
>>>>                        py::return_value_policy::reference_internal)
>>>>                   .def_property_readonly("size", &CameraConfiguration::size)
>>>>                   .def_property_readonly("empty", &CameraConfiguration::empty)
>>>> +               .def_readwrite("sensor_config", &CameraConfiguration::sensorConfig)
>>>>                   .def_readwrite("transform", &CameraConfiguration::transform);
>>>>
>>>>           pyCameraConfigurationStatus
>>>> --
>>>> 2.39.2
>>>>
>>