#216 iteratefdtd_matrix: Updates for the main loop

[willGraham01]

Closes #216 | Continues #63 |

Diff is large because .clang-format is running over all touched files. #222 merged so now the diff is manageable.

Source class changes:

• Now tracks whether the source data is empty through the private variable no_data_stored. This is safer than relying on comparing the data members real and imag to nullptr, since these are public members and so could be accidentally modified in the codebase. Remove MATLAB dependency #70 May change this ethos
• Has an access method is_empty() that can be used to access no_data_stored.

ObjectFromInfile changes:

• Has a fetch-method all_sources_are_empty() WDWYTID.

SimulationManager changes:

• post_processing() method now only runs the normalisation lines if at least one of the Source inputs is non-empty, as required.

source_term_updates.cpp

New file that breaks off the aforementioned Source update equations from the main loop code, and places them into functions. Also involves a logic update and better variable scoping.

• update_{IJK}source_terms_steadystate() handles the updates to the E_s, H_s, J_c, J_s (as appropriate) in steadystate simulations.
• update_source_steadystate() serves to unify the update equations, preserving the general structure of the updates whilst using switches to account for differences between the I,J,K cases.

TODO:

• E_s steadystate update equations
• E_s pulsed update equations
• H_s steadystate update equations
• H_s pulsed update equations
• Resolve flagged issues below

Issue Resolutions Required

• Impliments the suggestion in Out-of-scope k variable usage when updating source terms. #225 - issue remains open for further investigation.
• This PR now closes array_ind scoping issues #226

[samcunliffe]

Diff is large because .clang-format is running over all touched files.

Should we, therefore, merge #222 as-is, to simplify(?) this diff? Then clang-tidy and/or cpplint can come in later pull requests?

[willGraham01]

Diff is large because .clang-format is running over all touched files.

Should we, therefore, merge #222 as-is, to simplify(?) this diff? Then clang-tidy and/or cpplint can come in later pull requests?

Yes - I've made #222 ready for review so it can go in. Then I can rebase/merge this on top of that.

[samcunliffe]

Sorry for the slow review!
Weirdly I reviewed almost all of this on Monday and... then just stopped(!?)

All looks very good to me.
SimulationManager is now a large class although that's a big improvement from a gigantonormous function.

[willGraham01]

Sorry for the slow review! Weirdly I reviewed almost all of this on Monday and... then just stopped(!?)

All looks very good to me. SimulationManager is now a large class although that's a big improvement from a gigantonormous function.

No worries 😃 However the changes based on your comments mean that GitHub wants a re-review. But all that's changed is me doing this and resolving the conflicts with main.

[samcunliffe]

GitHub wants a re-review

Hm. Interesting. Was browsing the PR rules when looking into automerge but I didn't think I changed anything.
(This is probably my fault.)