Expose TaskContactForceEquality python binding

[danielcostanzi18]

Hello everyone,

First off, thanks a lot for the great work on TSID!
I want to use TSID to control the force at the end effector of a manipulator, to do so I want to use the task TaskContactForceEquality. The problem is that it's python binding is not exposed.
I tried to naively expose it with these changes.

When I try to instantiate the task I get the following error:

ArgumentError: Python argument types in
TaskContactForceEquality.init(TaskContactForceEquality, str, RobotWrapper, float, Contact6d)
did not match C++ signature:
init(_object*, std::__cxx11::basic_string<char, std::char_traits, std::allocator > name, tsid::robots::RobotWrapper {lvalue} robot, double dt, tsid::contacts::ContactBase {lvalue} contact)

It seems that the problem lies in the type of contact that I am passing, which is of type tsid.tsid_pywrap.Contact6d. From the C++ signature, it appears that the constructor expects tsid::contacts::ContactBase instead.

Could anyone shed light on how I might resolve this mismatch or if there's an alternative way to expose the correct binding for Contact6d?

If you want to reproduce the issue you can launch the example that I prepared exercizes/ex_5_ur5_force_task.py.

Thanks in advance for your help!

Best regards,
Daniel

[andreadelprete]

Hi @danielcostanzi18
it seems to me the problem is that in Python Contact6d is not recognized as a ContactBase, even though it is. Indeed if you take a look at how other Python bindings have been implemented (such as this one) you need to implement a different method for each possible child class of the input argument. In your case, you would need a constructor for each possible child of ContactBase (such as Contact6d).

[danielcostanzi18]

Hi @andreadelprete
Thanks a lot for your help, I'm now able to instantiate the task by passing a Contact6d.
These are the last changes.
I'd like to open a PR, but first would probably be better to have a working example, what do you think?
For that matter, I'd open another issue.

[andreadelprete]

Yes that would be ideal thanks!