Bamboost

Simulation

Every collection is made off many simulations. On the individual simulation level, bamboost offers an extensive API to make writing and reading of numeric data as easy as possible.

A big part of bamboost is its data model for simulation data. Every simulation has:

  • A primary hdf5 file named data.h5 This is the core of each simulation. It holds it's metadata, parameters, and data.
  • (optional, but likely) An executable that computes the result
  • (optional, but likely) A bash script that usable to start the executable (or to submit it to a HPC task scheduler)
  • (optional) Any other files relevant to the user

Primary hdf file

The bamboost interface for simulations is built around h5py.

The file data structure is illustrated below.

Placeholder...

Mutability

By default, the hdf file is immutable. This means that no bamboost code will edit the file. However, the file itself is still mutable, so be careful when using pure h5py.

To allow editing with bamboost, you must obtain a SimulationWriter:

sim_editable = sim.edit()