Normalize Data by Source Spectrum

[tylerflex]

Currently, the SimulationData object loads the raw field, flux, and mode data.
This PR provides a way to normalize frequency-dependent data by the spectrum of one of the original sources.
This eliminates the need to run a separate normalization simulation in some cases.

Main changes to the API:

• FreqData (frequency-dependent monitor data) has a .normalize(source_amps) abstract method, that changes their .values in place given some source amplitude spectrum.
• ScalarFieldData, FluxData, and ModeData implement .normalize(source_amps) methods for their respective data types (depending on whether they are field-like or power-like data).
• sim_data_normalized = SimulationData.normalize(normalize_index) normalizes all of the FreqData in the SimulationData using the respective .normalize() functions defined above. normalize_index indexes into the Simulation.sources list to grab the source time-dependence to normalize by. A copy is created, as to preserve the original data in case one wants to normalize by a different source later.
• web.load() adds a normalize_index argument to allow users to normalize upon loading the data into SimulationData objects.
• This normalize_index is propagated to all places in web API and web container that call web.load().

Tests show Flux==1 to 1e-4 tolerance for plane wave.

[momchil-flex]

Looks good and passes both my PlaneWave and GaussianBeam unit flux tests. Fine by me to merge.