Fix nan2, isVis, and other errors #88
Conversation
|
@djarecka do you want to take a look at it here? if everything is good, we can merge and then I move on to |
|
don't the abcd and hbcd csv's have multiple languages? |
|
@yibei - I will take a closer look later, but quick one thing. when I was debugging last week, I noticed that using |
|
@satra but we only have the english version of HBCD, like there are no multi-language in one CSV file @djarecka I guess you mean the following where (strip=True) will remove |
I believe |
|
@djarecka i set |
| assert not errors_list, f"Errors: {errors_list}" | ||
| print("No errors, but found warnings: ", warnings_list) | ||
| # More informative assertion | ||
| real_errors = [err for err in errors_list if err is not None] |
There was a problem hiding this comment.
the errors should never be None. That happened because a combination of two things:
- the new errors that were introduced in the previous PR fix nan #81
- and an error in the test: instead using
print_return_msgthere isprintfunction added to the error list in one place, so there areNones instead of the message itself. I fixed it in fixing test_rs2redcap_redcap2rs and reverting some changes to converters #86, see here
So unfortunately it means that there are still errors, with isVis (you will see specific message after the test fix)
|
@djarecka I copied your here is our output, yes our biggest difference is in
|
|
if you changed the logic of setting |
based on the issue we've discussed here #81 #86 I did
dsm_5_parent_guardian_rated_level_1_crosscutting_s, which causes some unexplainable issue. i changed it tomainthe original:
our conversion:
this is a different issue—language, which is involved in both
reproschema2redcap.pyandredcap2reproschema.py. I suggest that we put it aside for a while until we have some redcap csv files that have mutli-language so that we can know how languages are set up there.