Reputation: 11
In PsychoPy v1.85.1 I am configuring the 'polygon' stimuli object. I am trying to set the height attribute via a csv conditions file. I get an error message saying "Invalid parameter. Single numbers are not accepted"
PsychoPy v1.85.1 on Windows.
In the polygon's pop-up UI 'size' box I enter: $height
In the csv-file I have a 'height' column. Each row has values such as: (1.5, 0)
PsychoPy gives the error message:
File "C:\Program Files (x86)\PsychoPy2\lib\site-packages\psychopy \visual\basevisual.py", line 1312, in pos
self.__dict__['pos'] = val2array(value, False, False)
File "C:\Program Files (x86)\PsychoPy2\lib\site-packages\psychopy\tools\arraytools.py", line 176, in val2array
raise ValueError(msg % str(length))
ValueError: Invalid parameter. Single numbers are not accepted. Should be tuple/list/array of length 2
I go to the arraytools.py file and find the relevant code snippet. I paste it into a Python notebook (although it is python 3.3) and add some print rows for debugging:
# Copied code snippet from
# C:\Program Files (x86)\PsychoPy2\lib\site-packages\psychopy\tools\arraytools.py
import numpy
def val2array(value, withNone=True, withScalar=True, length=2):
"""Helper function: converts different input to a numpy array.
Raises informative error messages if input is invalid.
withNone: True/False. should 'None' be passed?
withScalar: True/False. is a scalar an accepted input?
Will be converted to array of this scalar
length: False / 2 / 3. Number of elements input should have or be
converted to. Might be False (do not accept arrays or convert to such)
"""
if value is None:
if withNone:
return None
else:
raise ValueError('Invalid parameter. None is not accepted as '
'value.')
value = numpy.array(value, float)
print ("value:", value) #I ADDED
print ("value.shape:", value.shape) #I ADDED
print ("numpy.product(value.shape):", numpy.product(value.shape)) #I ADDED
if numpy.product(value.shape) == 1: #MY COMMENT: WHY DOES THIS EVALUTE TRUE?
if withScalar:
# e.g. 5 becomes array([5.0, 5.0, 5.0]) for length=3
return numpy.repeat(value, length)
else:
msg = ('Invalid parameter. Single numbers are not accepted. '
'Should be tuple/list/array of length %s')
raise ValueError(msg % str(length))
elif value.shape[-1] == length:
return numpy.array(value, float)
else:
msg = 'Invalid parameter. Should be length %s but got length %s.'
raise ValueError(msg % (str(length), str(len(value))))
I test it by entering a value and then run the function.
# Run the function
value = (1.5,0.0)
val2array(value, False, False, length =2)
Results below. Seems to work fine:
value: [ 1.5 0. ]
value.shape: (2,)
numpy.product(value.shape): 2
Out: array([ 1.5, 0. ])
Thank you Michael. It seems the input value becomes a unicode string which the numpy arrary function cannot convert
print "position: ", position
print "type(position): ", type(position)
print "numpy.array(position, float): ", np.array(position, float)
#Results:
position: (0, 0.5)
type(position): <type 'unicode'>
numpy.array(position, float):
Traceback (most recent call last):
File "C:\Users\nikla\Documents\PsychoPy\test2.py", line 127, in <module>
print "numpy.array(position, float): ", np.array(position, float)
ValueError: could not convert string to float: (0, 0.5)
Any idea what I am doing wrong?
Upvotes: 0
Views: 853
Reputation: 11
The problem was fixed by entering the attribute values (in the GUI pop-up) using square brackets:
[1.5, 0]
Upvotes: 1
Reputation: 5683
The error has to do with the position attribute rather than the height. Do you set the position to be a scalar? It has to be a coordinate. Try writing the position in your csv file as (1.5, 0.0)
and then enter $eval(position)
in Builder (or whatever you called the position column).
Upvotes: 0