Renal reg #781
Renal reg #781
Conversation
This adds with_record_in_UKRR() to patients.py for the purpose of passing through arguments.
These should fail.
rebkwok
left a comment
rebkwok
left a comment
There was a problem hiding this comment.
Looks good so far - just a few minor comments
cohortextractor/patients.py
Outdated
| def with_record_in_UKRR( | ||
| # picks dataset held by UKRR | ||
| from_dataset, | ||
| returning, |
There was a problem hiding this comment.
pretty much everywhere else in patients.py we use keyword args only - I'm assuming that's to prevent users putting positional args in the wrong order
There was a problem hiding this comment.
I'm not sure I completely understand this comment. Do you mean change the order of them or remove? We need to be able to identify the dataset in the UKRR somehow.
There was a problem hiding this comment.
I just meant to make them keywords instead of positional args. It means users can't mix up the order and specify returning where we think we're getting from_dataset.
Like:
def with_record_in_UKRR(
# picks dataset held by UKRR
from_dataset=None,
returning="binary_flag",
...
I think elsewhere we generally default to binary_flag for returning. from_dataset could have a default if it makes sense to default to one of them, or None and we raise an exception if they don't provide it
tests/test_tpp_backend.py
Outdated
|
|
||
| assert_results( | ||
| study.to_dicts(convert_to_strings=False), | ||
| renal_registry_2019_prev=[1, 0] |
There was a problem hiding this comment.
I think binary_flag returns strings (but the test output will tell us that, I guess :) )
| ICHD, HHD, HD, PD, Tx | ||
| * "treatment_modality_prevalence" - the treatment modality from the prevalence data | ||
| * "latest_creatinine" - most recent creatinine held by UKRR | ||
| * "latest_egfr" - most recent eGFR held by UKRR |
There was a problem hiding this comment.
If it makes more sense, we can map these things to more user-friendly labels (same for the dataset names). I'm not sure if it's helpful here, but some of the other methods do (e.g. patients_admitted_to_hospital).
There was a problem hiding this comment.
Let's talk about this
cohortextractor/patients.py
Outdated
| * "latest_creatinine" - most recent creatinine held by UKRR | ||
| * "latest_egfr" - most recent eGFR held by UKRR | ||
| on_or_before: date of interest as a string with the format `YYYY-MM-DD`. Filters results to measurements | ||
| on or before the given date (as defined by `date_filter_column`). |
There was a problem hiding this comment.
copy/paste error (and in the next 2 args)? There's no date_filter_column
cohortextractor/tpp_backend.py
Outdated
| on_or_before=None, | ||
| on_or_after=None, | ||
| between=None, | ||
| include_date_of_match=False, |
| # Date filtering: date limits | ||
| on_or_before=None, | ||
| on_or_after=None, | ||
| between=None, |
There was a problem hiding this comment.
We only need between here - process_covariate_definitions.handle_time_period_options converts any on_or_before or on_or_after args into a between one before it gets to the backend.
cohortextractor/tpp_backend.py
Outdated
| elif returning == "latest_egfr": | ||
| column_definition = "eGFR_ckdepi" | ||
| else: | ||
| column_definition = returning |
There was a problem hiding this comment.
This could be nicer as another mapping:
column_mapping = {
"binary_flag": 1,
"treatment_modality_start": "mod_start",
...
}
column_definition = column_mapping.get(returning)
And then we can also check and raise an exception if the returning value isn't one we expected
tests/test_tpp_backend.py
Outdated
| renal_centre="AD100", | ||
| rrt_start="2019-01-01", | ||
| mod_start="RT", | ||
| mod_prev="NULL", |
There was a problem hiding this comment.
I'm not entirely confident about this, but I think you want None here, and sqlalchemy will make it a null in the db. (Unless nulls in the real data are represented by the string "NULL".)
rebkwok
left a comment
rebkwok
left a comment
There was a problem hiding this comment.
I left a few minor suggestions/questions, but this looks good 👍
THere's something formatting-related that black is complaining about atm
On my part
Start of adding the renal registry to the backend.