Parameters

Every simulation stores its input parameters inside the parameters group of data.h5. The Parameters object behaves like a Python dictionary, but transparently reads from and writes to the HDF5 file.

Reading parameters

Access sim.parameters on any Simulation object to get a dictionary-like view of all stored parameters.

from bamboost import Collection

coll = Collection(uid="315628DE80")
sim = coll["my-simulation"]

# Dictionary-style access
Re = sim.parameters["Re"]
dt = sim.parameters["dt"]

# Nested access with dot notation
radius = sim.parameters["body.radius"]

# Iterate over all parameters
for key, value in sim.parameters.items():
    print(key, value)

Writing parameters

Parameters can only be written via a SimulationWriter.

sim = coll["my-simulation"].edit()
sim.parameters["Re"] = 1000
sim.parameters["dt"] = 5e-3

# Update multiple at once
sim.parameters.update({"nu": 1e-3, "body": {"E": 1e6, "radius": 0.5}})

numpy arrays are stored as HDF5 datasets inside the parameters group. All other types (scalars, strings, lists) are stored as HDF5 attributes.

Parameters in the collection DataFrame

When querying the collection with coll.df, every parameter key becomes a column. Nested parameters appear as flattened dot-separated column names (body.E, body.radius).

coll.df[["name", "Re", "dt", "body.E"]]

Parameters as collection filter keys

Use coll.k["param_name"] to build filter expressions for coll.filter(...).

filtered = coll.filter(
    coll.k["Re"] >= 500,
    coll.k["body.radius"] < 1.0,
)

See Collection for more details on filtering.