This README is updated 29 May 2022.

A thorough introduction to FHIR in the NIH Cloud Interoperability project is provided in a jupyter notebook last edited in July 2020.

In the RStudio console, in May 2022, use

pip install pyAnVIL==0.1.1rc2
pip install git+https://github.com/smart-on-fhir/client-py#egg=fhirclient==4.0.0

to obtain necessary infrastructure.

The R package at github.com/vjcitn/AnvBiocFHIR uses basilisk to pin down all details of python infrastructure used to interface to the FHIR services in AnVIL.

Ensure that the .Renviron file in the home folder has a line PIP_USER=false.

Verify that the setting of PIP_USER is "false" using Sys.getenv("PIP_USER").

Install the package:

BiocManager::install("vjcitn/AnvBiocFHIR")

Instantiate the python infrastructure with

x = try(AnvBiocFHIR::abfhir_demo()) 

This will take some time to construct a conda environment.

Evaluate x to see a list of two Module objects produced by reticulate.

library(AnvBiocFHIR)
example(connect_smart)

On 29 May 2022 the result is

> unlist(lapply(res[1:5], function(x) x$id))
[1] "AnVIL-CMG-Broad-Muscle-Myoseq-WES"       
[2] "AnVIL-CMG-Broad-Orphan-Estonia-Ounap-WGS"
[3] "AnVIL-CMG-Broad-Brain-Gleeson-WES"       
[4] "AnVIL-CMG-Broad-Muscle-Topf-WES"         
[5] "AnVIL-CCDG-Broad-CVD-AFib-Duke-WGS"