episimlab.utils package

Submodules

episimlab.utils.datetime module

class episimlab.utils.datetime.IntPerDay

Bases: object

Provide an interval per day int_per_day

int_per_dayattr.Attribute

No description given

Variable properties:

type : variable
intent : out
global name : int_per_day
dimensions : ()

run_step(step_delta)

episimlab.utils.datetime.discrete_time_approx(rate, timestep)

Parameters

rate – daily rate
timestep – timesteps per day

Returns

rate rescaled by time step

episimlab.utils.datetime.dt64_to_day_of_week(dt64: numpy.datetime64) → int: Convert a datetime to integer corresponding to day of the week. Monday == 0, Sunday == 6.

episimlab.utils.datetime.get_int_per_day(step_delta: numpy.timedelta64) → float: Given a time delta, calculate the number of intervals per day.

episimlab.utils.plot module

episimlab.utils.plot.plotter(flavor='mpl', log_dir='./logs', log_stub=None, plotter_kwargs={})

TODO WORK IN PROGRESS

Decorates func with function that plots DataArray. This function returns a decorator, so use like:

@plotter() def my_func():

return xr.DataArray()

episimlab.utils.plot.visualize_compt_graph(gph: networkx.classes.graph.Graph, path: Optional[str] = None, mpl_backend: Optional[str] = None, pos=None, default_edge_color: str = 'black', default_node_color: str = 'black', node_size: int = 1000, font_size: int = 16, font_weight: str = 'bold', font_color: str = 'white', **kwargs): Visualize compartment graph gph using matplotlib. Saves figure to a path if specified, e.g. ‘my_compt_graph.svg’.

episimlab.utils.plot.xr_plot(data_array, sel={}, isel={}, timeslice=slice(0, 100, None), sum_over=['risk_group', 'age_group']): Uses DataArray.plot, which builds on mpl

episimlab.utils.profile module

episimlab.utils.profile.profiler(flavor='wall_clock', log_dir='./', log_stub=None, show_prof=False, cumulative=False)

Decorates func with Dask memory and thread profiling. This function returns a decorator, so use like:

@profiler() def my_func():

pass

episimlab.utils.rng module

episimlab.utils.rng.get_rng(seed: numbers.Number) → numpy.random._generator.Generator: Generates a np.random.Generator instance from a seed.

episimlab.utils.tags module

episimlab.utils.tags.tag(tag_name): Decorator factory that returns a tag_decorator, which adds string tag_name`s to the decorated object’s `TAGS attribute.

episimlab.utils.variable module

episimlab.utils.variable.any_negative(val, raise_err=False) → bool: Checks if there are any negative values in the array. Accepts ndarray and DataArray types.

episimlab.utils.variable.clip_to_zero(val): numpy.clip that accepts ndarray and DataArray types.

episimlab.utils.variable.coerce_to_da(proc, name: str, value, coords: Optional[dict] = None) → xarray.core.dataarray.DataArray: Given a variable with name and value defined in process proc, retrieve the variable metadata and use it to coerce the value into an xarray.DataArray with the correct dimensions and coordinates. Returns value if variable is scalar (zero length dims attribute), DataArray otherwise.

episimlab.utils.variable.edge_weight_name(u, v) → str: Attr name when looking for edge weight between nodes u and v.

episimlab.utils.variable.fix_coord_dtypes(da: xarray.core.dataarray.DataArray, max_len: Optional[int] = None) → xarray.core.dataarray.DataArray: Changes coords with object dtype to unicode, e.g. <U5, where 5 would be max_len in this case. Workaround for missing object_codec for object array.

episimlab.utils.variable.get_var_dims(process, name: str) → tuple: Given process-wrapped class process, retrieve the dims metadata attribute for variable with name.

episimlab.utils.variable.group_dict_by_var(d: dict) → dict: Given a dictionary keyed by 2-length tuples, return a dictionary keyed only by the second element of the tuple.

episimlab.utils.variable.suffixed_dims(da: xarray.core.dataarray.DataArray, suffix: str, exclude: Optional[list] = None) → dict: Generates a dictionary for suffixing DataArray or DataFrame dimensions. Compatible with methods such as DataFrame.rename and DataArray.rename.

episimlab.utils.variable.trim_data_to_coords(data: numpy.ndarray, coords: list) → numpy.ndarray: Given a data array, trim the data to match the shape given by list of 2-length tuples coords, formatted like [(‘dim0’, range(10)), (‘dim0’, range(10)), …]. Be sure to check that order of dims in data is same as order of dims in coords.

episimlab.utils.variable.unsuffixed_dims(da: xarray.core.dataarray.DataArray, suffix: str) → dict: Like suffixed_dims, but reverses the renaming transformation.