Add dpd contribution to pressure and pressure tensor

[mebrito]

Fixes #5030

Description of changes:

[jngrad]

Wouldn't it be easier to make DPD stress part of the pressure calculation in the core, like in #5043? With the currently proposed changes, ESPResSo scripts written in C++ and Python disagree about the total pressure.

[mebrito]

I also thought on doing it in this alternative way, but since energy and pressure use the same class in the core for calculation, that would lead to an extra meaningless "dpd" key in the energy dictionary.

[jngrad]

That would have been fine. Other features also have zero contribution in several fields, for example the cold Drude thermostat doesn't contribute to the energy.

[jngrad]

Wouldn't it be easier to instead extend the existing DPD pressure test?

espresso/testsuite/python/dpd.py

Lines 395 to 399 in 5c7ae80

	dpd_obs = espressomd.observables.DPDStress()
	obs_stress = dpd_obs.calculate()
	np.testing.assert_array_almost_equal(np.copy(dpd_stress), stress)
	np.testing.assert_array_almost_equal(np.copy(obs_stress), stress)

One line to get the DPD pressure via system.analysis.pressure_tensor()["DPD"] and one line to check it's equal to -stress (pressure should be the negative of the stress).