[open] Replacement function for numpy.opentxt in Droid backend/Android environment?

jnaveblod

Hello!

I'm a coding novice and trying to import an experiment I programmed in a different proprietary runtime over to OpenSesame, as I'm interested in a portable, tablet version of my experiment. I've worked and worked on it, and got it to work just fine on my computer - but it won't work on my tablet.

I should have looked closer at the API for the Droid backend, but I've been hobbling together my code based on some of the discussions here in the forum. Namely, I have multiple lines of text I want to present in order to the participants in each sequence - my experiment is intended to be used with children as well as adults, and stories and varied pictures are required to keep my participants' interest.

Rather than creating the lists of pictures, text lines, and sounds inside OpenSesame, I was using the method found here ( http://www.cogsci.nl/forum/index.php?p=/discussion/502/solved-reading-in-stimuli-list/p1 ) and here ( http://forum.cogsci.nl/index.php?p=/discussion/469/solved-selective-sampling-with-a-ratio-help-please ) and using text files read in through the NumPy modules. My code at the beginning of the experiment looks like this:

path_instructionslist = exp.get_file("bmsaa_kids_instructions.txt")
path_instructionspicslist = exp.get_file("bmsaa_kids_instructions_pics.txt")

...

global instructionslist, instructionspicslist

...

import numpy as np
instructionslist = np.loadtxt(path_instructionslist, dtype = str, delimiter = ";")
instructionspicslist = np.loadtxt(path_instructionspicslist, dtype = str, delimiter = ";")

and then later, before the relevant blocks, an inline script:

instructionslist = list(instructionslist)
instructionspicslist = list(instructionspicslist)

instructions_text = instructionslist.pop(0)
instructions_pic = instructionspicslist.pop(0)

exp.set("instructions_text", instructions_text)
exp.set("instructions_pic", instructions_pic)

It works beautifully on my windows desktop when I'm testing it (as NumPy is in included in the Windows environment), but when I attempt to run it on my Samsung tablet, it crashes out (well, exits gracefully without a word), and my traceback is that NumPy doesn't exist to be called in the Android environment. Whoops!

I've been desperately looking through the Python 2.7 documentation, but can't figure out what built-in Python function I can call to mimic the functionality of NumPy's loadtxt. I've tried the basic 'open()' -

instructionslist = open(path_instructionslist)
instructionspicslist = open(path_instructionspicslist)

in place of the 'np.opentxt(path_instructionslist)' in my first example, but then I get an error/traceback that the soundfile I want to call first doesn't exist, as it's added an extra bit of whitespace to my variable name. E.g. instead of 'silence' I get 'silence ' - so while 'silence.wav' exists, 'silence .wav' does not, hence the traceback.

What should I add to my code/what function can I call to get my program to work again? I'll keep poring over the documentation, but any and all help would be appreciated. Thank you!

jnaveblod

Okay - I've done a little googling and checked around, and this is what I came up with - problem is, while it should work (as far as I can tell), it still doesn't. The first bit is still the same; I've changed after I globally define my variables to read as thus:

instructionslist = [line.rstrip('\n') for line in path_instructionslist]
instructionspicslist = [line.rstrip('\n') for line in path_instructionspicslist]

and so on and so forth. I've checked my code in PyCharm; when I run

path_instructionssoundslist = open('bmsaa_kids_instructions_sounds.txt', 'r')
instructionssoundslist = [line.rstrip('\n') for line in path_instructionssoundslist]
print type(instructionssoundslist)
print instructionssoundslist
for line in instructionssoundslist:
    print line

I get what I'm expecting; my output looks like this:

<type 'list'>
['silence', 'silence', 'silence', 'silence', 'silence', 'silence', 'silence', 'silence', 'silence', 'music_alone', 'silence', 'click_alone', 'practice_synchronous01', 'silence', 'silence', 'silence', 'silence', 'silence', 'silence', 'silence', 'silence']
silence
silence
silence
silence
silence
silence
silence
silence
silence
music_alone
silence
click_alone
practice_synchronous01
silence
silence
silence
silence
silence
silence
silence
silence

So far so good - no extra whitespace at the end, the variable type IS list, and it seems to be doing what I ask - those are the values in the text file.

The inline_script at the beginning of the first instructions loop/sequence is such:

instructionslist = list(instructionslist)
instructionspicslist = list(instructionspicslist)
instructionssoundslist = list(instructionssoundslist)

instructions_text = instructionslist.pop(0)
instructions_pic = instructionspicslist.pop(0)
instructions_sound = instructionssoundslist.pop(0)

exp.set("instructions_text", instructions_text)
exp.set("instructions_pic", instructions_pic)
exp.set("instructions_sound", instructions_sound)

I can probably take out the first three lines, as they're redundant by this point, right? That's not the main question though.

The text and pictures on the canvas item in the sequence are dictated by [instructions_pic] and [instructions_text], and the sound in the sampler is dictated by [instructions_sound].

Relevant sampler script:

set sample "[instructions_sound].wav"

Now when I attempt to quick-run my experiment, though, I get the following traceback:

>>> 
Starting experiment as ExperimentProcess-33
openexp.sampler._legacy.init_sound(): sampling freq = 44100, buffer size = 1024
openexp.sampler._legacy.init_sound(): mixer already initialized, closing
experiment.init_log(): using 'D:/HDD Documents/2014-2015/BMSAA\quickrun.csv' as logfile (utf-8)
experiment.run(): experiment started at Fri Jun 19 19:40:04 2015

Failed to load sample in sampler "instructionssounds": 
openexp._sampler.legacy.__init__() the file 'D.wav' does not exist


>>> 

I cannot figure out for the life of me where it's getting "D" as the value of instructions_sound. Adding a line to PyCharm asking it to return the first item in the instructionssoundslist list gives me 'silence' (what I'd expect), but that's apparently NOT what OpenSesame is returning. I'm at my wits' end right now - according to PyCharm, my code SHOULD work. According to OpenSesame, it doesn't. Any ideas on what to tweak would be greatly appreciated!

eduard

Hey,

This sounds odd. Have you tried to use your other two variables (text and image) in a similar way, i.e. in a sketchpad? If it works for them, it could be a problem of the sampleritem. I couldn't try it myself, because I don't have the wav-files, but displaying the words in that way seems to work. If you could share your audio files, I can try it again and see whether I can reproduce this error.

I can probably take out the first three lines, as they're redundant by this point, right?

Yes, I think so.

Thanks,

Eduard

sebastiaan

I cannot figure out for the life of me where it's getting "D" as the value of instructions_sound.

I didn't fully read this discussion (so this is a bit of a hunch), but this comment struck me, and I think something like the following is going on.

• You have a path, for example: D:/HDD Documents/2014-2015/BMSAA
• You cast this to a list, so you get a list of separate characters, the first of which is D.
• You pop() the first character (i.e. D) from this list, and set that as the variable instructions_sound.

Could it be something like this?