Psychopy: MultiStairHandler method saveAsExcel creates too many files

[janfreyberg]

Hi,

This is a minor issue but I have a MultiStairhandler with 4 conditions (or 4 interleaved staircases).

When I'm done with the staircases, I call staircases.saveAsExcel. This produces 2 .xlsx files, with the first staircase being in the first spreadsheet, and the three subsequent staircases being in the second spreadsheet. It looks like this:

As you can see, the staircase labels are all unique. Only one of the staircases is in a separate .xlsx file.

For reference, I am using:
psychopy version 1.82.1
python version 2.7.11 (via anaconda 4.0.0, 32 bit)

[hoechenberger]

Does this problem still occur with the latest version of PsychoPy?

[janfreyberg]

Hi,

thanks for taking a look at this. However, there seem to be a whole host of new issues with this method.

With PsychoPy 1.84.2 (and openpyxl 2.4.0), I am now getting an ImportError when trying to call saveAsExcel:
ImportError: openpyxl is required for saving files in Excel (xlsx) format, but was not found.

I did a little digging, and found that this seems to stem from line 29 in data.py:
from openpyxl.cell import get_column_letter

as far as I can tell, get_column_letter sits in openpyxl.utils.cell:
http://openpyxl.readthedocs.io/en/default/api/openpyxl.utils.cell.html#openpyxl.utils.cell.get_column_letter

Now, if I change line 29 to import from openpyxl.utils.cell instead, I don't get the ImportError anymore. However, I get a different error instead:

File "[local python path]/data.py", line 3326, in saveAsExcel:
ew = ExcelWriter(workbook=wb)
TypeError: __init__() takes exactly 3 arguments (2 given)

I'll do some more digging to see what arguments are missing here, but it seems some update to openpyxl has broken this method?

Cheers,
Jan

[janfreyberg]

Just to add to that, ExcelWriter seems to require two arguments, workbook and archive:
http://openpyxl.readthedocs.io/en/default/_modules/openpyxl/writer/excel.html#ExcelWriter

I'm not sure what archive is, though.

[hoechenberger]

Hi @janfreyberg, thanks for your feedback. Could you please provide a minimal code example to reproduce the problem?

[peircej]

Yes, it looks like a change in the openpyxl APA. It seems we don't need to use the ExcelWriter explicitly at all though. From the docs here:
http://openpyxl.readthedocs.io/en/default/tutorial.html#saving-to-a-file
it looks like we can forget of any mention of it and just do workbook.save(filename) rather than

ew = ExcelWriter(workbook); 
ew.save(filename)

[janfreyberg]

I commented out the use of ExcelWriter and instead changed it to wb.save(filename) like you suggested. This fixes the problems with writing to excel, as it now creates excel files without error again.

However, it looks like the issue with multiple Excel files being created is still present.

This is a minimal example that reproduces the issue:
https://gist.github.com/janfreyberg/fece6b94493e66f5d98644fcc9ff9037

[hoechenberger]

Thanks, I will look into this later today.

Sent from my phone, please excuse brevity and erroneous auto-correct.

On 14 Oct 2016, at 10:29, Jan Freyberg [email protected] wrote:

I commented out the use of ExcelWriter and instead changed it to wb.save(filename) like you suggested. This fixes the problems with writing to excel, as it now creates excel files without error again.

However, it looks like the issue with multiple Excel files being created is still present.

This is a minimal example that reproduces the issue:
https://gist.github.com/janfreyberg/fece6b94493e66f5d98644fcc9ff9037

—
You are receiving this because you were assigned.
Reply to this email directly, view it on GitHub, or mute the thread.

[hoechenberger]

@janfreyberg I've narrowed this down to a simple bug in tools.fileerrortools.handleFileCollision(). Will propose a fix tonight.

[hoechenberger]

Discovered and fixed another bug on the way. See #1273.

Could you please open another issue for the openpyxl compatibility thingy? Thank you!

[janfreyberg]

Brilliant, thanks! I've created a new issue.

[hoechenberger]

Awesome, thank you very much!