You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
This commit was created on GitHub.com and signed with GitHub’s verified signature.
* added other domains to generate_included_concept_list
* added alias to the join
* added observation and device_exposure to spark_utils
* added cache_events when building NestedCohortBuilder
* used datetime to filter for the ehr records
* fixed the inded datetime for cabg
* fixed index date for atrial_fibrillation.py
* fixed index_date for coronary_artery_disease.py
* fixed the index date for ischemic_stroke.py
* fixed the index date for last_visit_discharged_home.py
* fixed the index date for type_two_diabietes.py
* fixed the person_id reference in the cohort definitions
* fixed the index dates for the patient outcome prediction tasks
* updated the index dates for the disease cohorts by using the condition_start_datetime
* updated the how the records are filtered
* fixed an alias for referencing condition_start_date
* removed filtering
* Revert "removed filtering"
This reverts commit 2dcb9137d49eec8eaf9828faf4e01acb64f42c2c.
* cache all ehr events to the disk for constructing cohorts
* added 0.1 second to the index date before we apply the filter to extract the features from the patient timelines
* updated the type two diabetes index date definition
* added a function to construct artificial visits for the problem list type of data
* added the logic for fixing the visit_occurrence_id in patient_events
* set visit_concept_id to 0 when the visit_occurrence_id is invalidated
* reconstruct visit_occurrence_ids for problem list events
* updated the logic for extracting features
* updated the index dates of the phenotype definitions
* added a timestamp format to to_timestamp
* convert the timestamp from utc to nyc
* removed time zone conversion
* set the index_date to visit_start_date + 23:59:00 for phenotypes if the condition_start_datetime occurred before the corresponding visit_start_datetime
* swtiched from date_add to the sql expression to keep the index_date as timestamp
* updated the test set up
* added optional caching for artificial visits
* Revert "updated the test set up"
This reverts commit e8962a8d4adcce8aeb47993fa5e33f6380c32d49.
* fixed the output of the cohort
* added an option to construct artificial visits for the problem list records that occur years ahead
* restored inpatient_att_type in generate_training_data
* updated the default value for date_filter: