Skip to content

combining module

Functions for combining arrays.

Combine functions combine two or more NumPy arrays using a custom function. The emphasis here is done upon stacking the results into one NumPy array - since vectorbt is all about brute-forcing large spaces of hyper-parameters, concatenating the results of each hyper-parameter combination into a single DataFrame is important. All functions are available in both Python and Numba-compiled form.

apply_and_concat function

apply_and_concat(
    ntimes,
    apply_func,
    *args,
    n_outputs=None,
    jitted_loop=False,
    jitted_warmup=False,
    execute_kwargs=None,
    **kwargs
)

Run apply_func function a number of times and concatenate the results depending upon how many array-like objects it generates.

apply_func must accept arguments i, *args, and **kwargs.

Set jitted_loop to True to use the JIT-compiled version.

All jitted iteration functions are resolved using JITRegistry.resolve().

Note

n_outputs must be set when jitted_loop is True.

Numba doesn't support variable keyword arguments.

apply_and_concat_each function

apply_and_concat_each(
    tasks,
    n_outputs=None,
    execute_kwargs=None
)

Apply each function on its own set of positional and keyword arguments.

Executes the function using execute().

apply_and_concat_multiple_nb function

apply_and_concat_multiple_nb(
    ntimes,
    apply_func_nb,
    *args
)

Run apply_func_nb that returns multiple arrays number of times.

Uses custom_apply_and_concat_multiple_nb().

apply_and_concat_none_nb function

apply_and_concat_none_nb(
    ntimes,
    apply_func_nb,
    *args
)

Run apply_func_nb that returns nothing number of times.

Uses custom_apply_and_concat_none_nb().

apply_and_concat_one_nb function

apply_and_concat_one_nb(
    ntimes,
    apply_func_nb,
    *args
)

Run apply_func_nb that returns one array number of times.

Uses custom_apply_and_concat_one_nb().

combine_and_concat function

combine_and_concat(
    obj,
    others,
    combine_func,
    *args,
    jitted_loop=False,
    **kwargs
)

Combine obj with each in others using combine_func and concatenate.

select_and_combine_nb() is resolved using JITRegistry.resolve().

combine_and_concat_nb function

combine_and_concat_nb(
    obj,
    others,
    combine_func_nb,
    *args
)

Numba-compiled version of combine_and_concat().

combine_multiple function

combine_multiple(
    objs,
    combine_func,
    *args,
    jitted_loop=False,
    **kwargs
)

Combine objs pairwise into a single object.

Set jitted_loop to True to use the JIT-compiled version.

combine_multiple_nb() is resolved using JITRegistry.resolve().

Note

Numba doesn't support variable keyword arguments.

combine_multiple_nb function

combine_multiple_nb(
    objs,
    combine_func_nb,
    *args
)

Numba-compiled version of combine_multiple().

custom_apply_and_concat_multiple_nb function

custom_apply_and_concat_multiple_nb(
    indices,
    apply_func_nb,
    *args
)

Run apply_func_nb that returns multiple arrays for each index.

custom_apply_and_concat_none_nb function

custom_apply_and_concat_none_nb(
    indices,
    apply_func_nb,
    *args
)

Run apply_func_nb that returns nothing for each index.

Meant for in-place outputs.

custom_apply_and_concat_one_nb function

custom_apply_and_concat_one_nb(
    indices,
    apply_func_nb,
    *args
)

Run apply_func_nb that returns one array for each index.

select_and_combine function

select_and_combine(
    i,
    obj,
    others,
    combine_func,
    *args,
    **kwargs
)

Combine obj with an array at position i in others using combine_func.

select_and_combine_nb function

select_and_combine_nb(
    i,
    obj,
    others,
    combine_func_nb,
    *args
)

Numba-compiled version of select_and_combine().

to_2d_multiple_nb function

to_2d_multiple_nb(
    a
)

Expand the dimensions of each array in a along axis 1.

to_2d_one_nb function

to_2d_one_nb(
    a
)

Expand the dimensions of the array along the axis 1.