MoleculeInserter implementations #270
Description
Currently, only RandomInserter is implemented, that places molecules at random positions in the simulation box (also allowing random rotations). Another useful strategy would be to place molecules at positions that are read from a file, e.g., xyz or gro.
In the yaml/json configuration, the inserter attributes are now inappropriately and confusingly part of the molecule attributes. They should be move to the insertmolecules section.
Insertion of atomic molecules is quite straightforward. The positions are read from a file line by line where atom names match. When the coordinates need to wrap around the PBC, a warning is issued.
Insertion of complex molecules can be done similarly. It is presumed that the order of atoms (after any unknown atoms are filtered out) corresponds to the topology of the molecule. If a mismatch in atom order is found, an error shall be raised. This also means that the structure of the molecule provided in the molecule configuration will not be used at all as the structure is described here directly by the atoms positions.
(In the future, an alternative description of molecule placement could be considered: COM position and orientation vector.)
Comments are welcome.
Related to #268.