Notes

General notes to self and TODOs

Optimizations

Notes on potential optimizations to implement.

Consider using ndarray

Currently a BurnMatrix is based on a MathJS SparseMatrix. Another option is the javascript implementation of an ndarray, which may or may not be faster for reading and writing to a BurnMatrix

Bypass resample function on initial return

Currently, TimeSteps are calculated and added to the Campaign.timesteps array. They are not resampled until the Campaign is ready to be sent to the front end. The resample function can be expensive on the CPU. However, for initial calculations, it can be avoided altogether. Within the iteration of Timesteps, Timestep.snapshots can be added to a Campaign.snapshots array based on whether or not they satisfy the same constraints used when using resample, i.e., only adding a snapshot if timestep.time <= latestSampleTime just after adding the Timestep to Campaign.timesteps. The Campaign will always need to keep track of all Timesteps, as if a user wants to view the Campaign with a different timescale. For example, if a user initially wants to view a Campaign with a timescale of 1 hour, we can push Timestep.snapshots to the Campaign.snapshots array every hour upon initial calculation. If they then want to view the same Campaign with a timescale of 1 day, we can then use the resample function against Campaign.timesteps to return that sampling.