SOP AoU-EHR-File-Check
All of Us Standard Operating Procedure
|
Purpose
|
To create a standardized method for sites to download and implement the latest version of the AoU-EHR-File-Check to use with the OMOP CDMv 5.3 during the 2021 Q3 submission cycle
|
|
Scope
|
The process applies to all program partner organizations that are responsible for sharing participant data (e.g., EHR) with the DRC. The recently updated file-check can be downloaded from a GCS bucket.
|
|
Materials
|
Requirements:
-
Python 3 (download from here and install)
-
pip (download get-pip.py and run python get-pip.py)
|
|
Background
|
While a submission from sites is not required this quarter, we are requiring that sites make a test submission with CDMv 5.3 using the new aou-ehr-file-check. This is required for all HPO sites during Q3 from 7/15 to 9/1.
|
|
Procedure
|
- Cloning Repo (use the master branch)
- Through GIT
- Or download Repo
-
(First time) Install requirements by running
pip install -r requirements.txt
- Run file validator
-
Update _settings.py and rename it to settings.py
- Create a folder and place all the EHR submission files to be validated into it
-
Set the "csv_dir" parameter in settings.py to the full path of the folder created above
-
Ensure the resources folder is also downloaded and is located in the same place as omop_file_validator.py
-
Execute the following at the command line:
python omop_file_validator.py
- Validation logic
-
File names must follow naming convention {table}.csv
-
table an OMOP CDM table listed in resources/omop
-
Files must be in CSV format (comma-delimited) as specified by rfc4180
-
Column names and types must follow the conventions in resources/omop
-
Date and timestamps formats must be compatible with bigquery
- There can be no blank lines in the csv file
- Unless all lines are blank
- File errors
- Errors will be stored in the files results.html and results.csv
- Errors need to be solved before uploading to the bucket
|
| |
|
References:
https://github.com/all-of-us/aou-ehr-file-check#readme