Refactor particle management

[jngrad]

Description of changes:

• particle state tracking is now fully encapsulated
  • helps with separation of concerns and reducing code duplication
• particle creation and particle moves are now carried out by 2 independent functions
  • this makes the reaction methods bugs throw an error instead of silently reacting particles of the wrong type (Electroneutrality breaks when combining Widom + cpH methods #4595)

[pm-blanco]

@jngrad Maybe I did not understand properly what you meant in this PR but it does not seem to me that it actually throws an error when reacting particles of the wrong type. I tested your PR with the script that I provide in #4595 and it still does not show any error when producing the bug.

[jngrad]

@jngrad Maybe I did not understand properly what you meant in this PR but it does not seem to me that it actually throws an error when reacting particles of the wrong type. I tested your PR with the script that I provide in #4595 and it still does not show any error when producing the bug.

Here is the output on my workstation:

jgrad@seraue:/work/jgrad/espresso-fork-PR/build-lama-release2$ mpirun -n 1 ./pypresso mwe_pablo.py 
step  0
Before widom: a: 1 H: 1 Ha: 1 Na: 2 Cl: 2
total charge  0.0
/work/jgrad/espresso-fork-PR/build-lama-release2/src/python/espressomd/reaction_methods.py:462:
FutureWarning: arguments 'reactant_coefficients' and 'product_coefficients' are deprecated and
are no longer necessary for the constant pH ensemble. They are kept for backward compatibility
but might be deleted in future versions.
  warnings.warn(warn_msg, FutureWarning)
Traceback (most recent call last):
  File "mwe_pablo.py", line 105, in <module>
    widom.calculate_particle_insertion_potential_energy(reaction_id=0)
  File "/work/jgrad/espresso-fork-PR/build-lama-release2/src/python/espressomd/reaction_methods.py", line 515, in calculate_particle_insertion_potential_energy
    return self.call_method(
  File "script_interface.pyx", line 159, in espressomd.script_interface.PScriptInterface.call_method
RuntimeError: Invalid local particle index entry.

In the end, the bug needs to be removed with your PR.

[pm-blanco]

That's strange, somehow I cannot reproduce it in my workstation. I get

mpirun -n 1 ../../espresso_jn/build/pypresso  bug_widom.py 
/home/pablo/espresso_jn/build/src/python/espressomd/reaction_methods.py:462: FutureWarning: arguments 'reactant_coefficients' and 'product_coefficients' are deprecated and are no longer necessary for the constant pH ensemble. They are kept for backward compatibility but might be deleted in future versions.
  warnings.warn(warn_msg, FutureWarning)
step  0
Before widom: a: 2 H: 2 Ha: 0 Na: 2 Cl: 2
total charge  0.0
After widom: a: 2 H: 1 Ha: 0 Na: 2 Cl: 2
total charge  -1.0

In the end, the bug needs to be removed with your PR.

it is quite right that we need to solve it in another PR but, just to avoid confusion, my current PR is just on the refactoring of the MC function to move particles. I was waiting for your green light on my idea of checking the bookeeping list each time the do_reaction function is called as you mentioned that you wanted to discuss it with Rudolf in my issue, since it comes with an associated performance loss.

Anyway, I will not oppose to merge this PR for this error message since I agree with you that we need to solve the bug in another PR anyway.