Stop copying Dirichlet csv rows and fix the warning's expected count

[drbergman]

The my-physicell counterpart of #72. Note up front that the two defects #72 fixes have no direct counterpart here: sync_substrate_dirichlet_activation, the fix_/unfix_substrate_at_voxel(s) family and find_existing_density_index are all new on dc-rework and don't exist on this branch, so there is no mesh-copying loop and no split error policy to unify. What does correspond is in the Dirichlet-from-file reader.

get_row_from_dirichlet_condition_csv took line and substrate_indices by value, copying a string and heap-allocating a vector for every row of the file; both are now const&. This is the same change #69 makes to the substrate reader's twin signature, and it touches a different function, so the two shouldn't collide.

The "wrong number of density data" warning in that function printed number_of_voxels() where it means number_of_densities() — Found is a count of substrate columns, and load_dirichlet_conditions_from_matlab's equivalent message already uses the density count. Worth knowing: that warning is unreachable today. In the header-less branch of load_dirichlet_conditions_from_csv, i is never incremented past the continue, so substrate_indices ends up empty, and the re-opened std::ifstream file shadows the outer one and dies at the end of the block, leaving the outer stream closed. A header-less DC csv therefore sets nothing at all, silently. That is the Dirichlet twin of the header-less path #69 revives for substrates, and it is still missing — happy to do it as a follow-up if you want it in this round.

[Copilot]

Copilot was unable to review this pull request because the user who requested the review has reached their quota limit.