Module bvmpc.bv_array_methods
Assorted helper functions to split timestamp numpy arrays up.
Expand source code
"""Assorted helper functions to split timestamp numpy arrays up."""
import numpy as np
def split_into_blocks(array, blocks=None, block_max=None, num_blocks=None):
"""
Split a sorted array into num_blocks.
Each block has a value of at most block_max
greater than the previous block.
Returns
-------
List:
A list of size num_blocks
"""
if blocks is None:
if (block_max is None) or (num_blocks is None):
raise ValueError(
"block_max and num_blocks must be passed to" +
"split_into_blocks if param blocks is not passed")
blocks = np.arange(
block_max + 0.001, block_max * num_blocks + 0.001, block_max + 0.001)
return np.array(np.split(array, np.searchsorted(array, blocks)))
def split_array(array, idxs):
"""
Split a one d array into two arrays based on idxs.
Returns
-------
(array[idxs], array[not idxs])
"""
if len(idxs) == 0:
return [], []
left = array[idxs]
ia = np.indices(array.shape)
not_indices = np.setxor1d(ia, idxs)
right = array[not_indices]
return (left, right)
def split_array_with_another(array, split_arr):
"""
Split a sorted one d array into two.
First array is the values directly after splt_arr.
Returns
-------
(array values after split_array, other values)
"""
idxs = np.searchsorted(array, split_arr)
idxs = idxs[idxs != len(array)]
idxs = np.unique(idxs)
return split_array(array, idxs)
def split_array_in_between_two(array, left, right):
"""
Split a sorted one d array into two.
One array is the values in between left and right
left and right should be the same size
Returns
-------
(array values between left and right, other values)
Raises
------
ValueError
If 'left' is not the same size as 'right'
"""
if len(left) != len(right):
raise(ValueError(
'left and right must have the same number of elements'))
good_idxs = []
for i, val in enumerate(array):
bigger = (left <= val)
smaller = (right >= val)
between = np.logical_and(smaller, bigger)
if between.any():
good_idxs.append(i)
return split_array(array, good_idxs)
def check_error_during_tone(presses, block_s, l_type=" "):
"""
Removes spurious responses generated by system during tone
Parameters
----------
presses: np.array
Lever responses to be checked
block_s: np.array
tone start timestamps + 5
Returns
-------
presses: np.array
Corrected lever responses
"""
for i, (a, b) in enumerate(zip(block_s - 5, block_s)):
tone_errors = np.where((presses > a) & (presses < b))
if np.size(tone_errors) > 0:
print("Removed {} spurious{}errors during tone {}.".format(
len(tone_errors), l_type, i))
presses = np.delete(presses, tone_errors)
return presses
if __name__ == "__main__":
array = np.array([i + 1 for i in range(9)])
tests = []
tests.append(split_into_blocks(array, 3, 3))
tests.append(split_array(array, [0, 1, 2]))
tests.append(split_array_with_another(array, [0.9, 3.9, 8.9]))
tests.append(split_array_in_between_two(array, [1, 8], [4, 9]))
for t in tests:
print(t)Functions
def check_error_during_tone(presses, block_s, l_type=' ')-
Removes spurious responses generated by system during tone
Parameters
presses:np.array- Lever responses to be checked
block_s:np.array- tone start timestamps + 5
Returns
presses:np.array- Corrected lever responses
Expand source code
def check_error_during_tone(presses, block_s, l_type=" "): """ Removes spurious responses generated by system during tone Parameters ---------- presses: np.array Lever responses to be checked block_s: np.array tone start timestamps + 5 Returns ------- presses: np.array Corrected lever responses """ for i, (a, b) in enumerate(zip(block_s - 5, block_s)): tone_errors = np.where((presses > a) & (presses < b)) if np.size(tone_errors) > 0: print("Removed {} spurious{}errors during tone {}.".format( len(tone_errors), l_type, i)) presses = np.delete(presses, tone_errors) return presses def split_array(array, idxs)-
Split a one d array into two arrays based on idxs.
Returns
(array[idxs], array[not idxs])
Expand source code
def split_array(array, idxs): """ Split a one d array into two arrays based on idxs. Returns ------- (array[idxs], array[not idxs]) """ if len(idxs) == 0: return [], [] left = array[idxs] ia = np.indices(array.shape) not_indices = np.setxor1d(ia, idxs) right = array[not_indices] return (left, right) def split_array_in_between_two(array, left, right)-
Split a sorted one d array into two.
One array is the values in between left and right left and right should be the same size
Returns
(array values between left and right, other values)
Raises
ValueError- If 'left' is not the same size as 'right'
Expand source code
def split_array_in_between_two(array, left, right): """ Split a sorted one d array into two. One array is the values in between left and right left and right should be the same size Returns ------- (array values between left and right, other values) Raises ------ ValueError If 'left' is not the same size as 'right' """ if len(left) != len(right): raise(ValueError( 'left and right must have the same number of elements')) good_idxs = [] for i, val in enumerate(array): bigger = (left <= val) smaller = (right >= val) between = np.logical_and(smaller, bigger) if between.any(): good_idxs.append(i) return split_array(array, good_idxs) def split_array_with_another(array, split_arr)-
Split a sorted one d array into two.
First array is the values directly after splt_arr.
Returns
(array values after split_array, other values)
Expand source code
def split_array_with_another(array, split_arr): """ Split a sorted one d array into two. First array is the values directly after splt_arr. Returns ------- (array values after split_array, other values) """ idxs = np.searchsorted(array, split_arr) idxs = idxs[idxs != len(array)] idxs = np.unique(idxs) return split_array(array, idxs) def split_into_blocks(array, blocks=None, block_max=None, num_blocks=None)-
Split a sorted array into num_blocks.
Each block has a value of at most block_max greater than the previous block.
Returns
List
A list of size num_blocks
Expand source code
def split_into_blocks(array, blocks=None, block_max=None, num_blocks=None): """ Split a sorted array into num_blocks. Each block has a value of at most block_max greater than the previous block. Returns ------- List: A list of size num_blocks """ if blocks is None: if (block_max is None) or (num_blocks is None): raise ValueError( "block_max and num_blocks must be passed to" + "split_into_blocks if param blocks is not passed") blocks = np.arange( block_max + 0.001, block_max * num_blocks + 0.001, block_max + 0.001) return np.array(np.split(array, np.searchsorted(array, blocks)))