demichie / elicipy

Expert. elicitation Python code
GNU General Public License v2.0
2 stars 4 forks source link

fatal error when only one target question #5

Closed GeoBen93 closed 1 month ago

GeoBen93 commented 4 months ago

Hello,

I have succesfully been running elicipy on a number of elicitations. However, on the latest elicitation with one question I recieve the following fatal error message:

The new directory C:\Users######\elicipy-main/ELICITATIONS/Ixchel_test/OUTPUT is created! STEP1: Reading questionnaire STEP2: Reading all answers to define ranges Merging of individual csv files Merging target The new directory ../TARGET_NEW is created! Saving target answers |██████████████████████████████████████████████████| 100.0% STEP3: Reading answers for group 0 Merging of individual csv files Merging target First Name Last Name C I_tot I_real w normW 0 name redacted 0 0 0 0 12.5 1 name redacted 0 0 0 0 12.5 2 name redacted 0 0 0 0 12.5 3 name redacted 0 0 0 0 12.5 4 name redacted 0 0 0 0 12.5 5 name redacted 0 0 0 0 12.5 6 name redacted 0 0 0 0 12.5 7 name redacted 0 0 0 0 12.5 Creating samples for questions C:\Users######\elicipy-main\tools.py:24: RuntimeWarning: invalid value encountered in divide percents = str_format.format(100 (iteration / float(total))) C:\Users######\elicipy-main\tools.py:25: RuntimeWarning: invalid value encountered in divide filled_length = int(round(bar_length iteration / float(total))) Traceback (most recent call last): File "C:\Users######\elicipy-main\elicipy.py", line 2917, in main(sys.argv[1:]) File "C:\Users######\elicipy-main\elicipy.py", line 1545, in main create_samples(group, n_experts, n_SQ, n_TQ, n_pctl, SQ_array, File "C:\Users######\elicipy-main\elicipy.py", line 934, in create_samples printProgressBar(j, n_SQ + n_TQ - 1, prefix=' ') File "C:######\elicipy-main\tools.py", line 25, in printProgressBar filled_length = int(round(bar_length * iteration / float(total))) ValueError: cannot convert float NaN to integer

I have added a placeholder 2nd question to the questionnaire and questionnaire responses, and this resolves the issue. Looks like it's just a problem with the calculation of the loading bar on the python console when there is only one question.

demichie commented 4 months ago

Hi, I've make a small change but I cannot test it because I'm with laptop only and traveling. I hope it works.

GeoBen93 commented 2 months ago

Hi, Sorry it's taken so long replying to this, really appreciate your assistance.

I have downloaded the updated code but I am unfortunately I am still getting the same error as above when eliciting only 1 target question.

Thanks

demichie commented 2 months ago

Hi! I am traveling now. Please can you send me a reminder by email to check this thing in a couple of weeks?

Thanks!

GeoBen93 commented 2 months ago

Absolutely, no problem. I'll send a reminder in a few weeks. Thanks!

demichie commented 1 month ago

Hi, Do you have both seed and target questions, or only one target with weights read from a file?

Mattia

demichie commented 1 month ago

Hi Ben, Can you test if now it works?

Mattia

GeoBen93 commented 1 month ago

Hi Mattia, I have just tested it and it is now working as expected with 1 target question only. Apologies I didn't get back to you before. Thanks for your help, Ben