This minor release contains fixes for two issues:

1. A requirement on a "request" python modules that was removed from pypi caused installing python requirements to fail - this requirement was removed.
2. Eliminated an issue where files with duplicate file names with different casings in spec examples caused issues on some platforms.

This release adds Go FHIR Proto definitions, as well as a Go FHIR parser and printer library with support for STU3 and R4. The initial release does not yet support FHIRPath validation in Go, but that will be coming soon. A usage example can be found in the companion repo, https://github.com/google/fhir-examples/blob/master/go/google/fhir_examples/parse_patients.go.

Other Changes

• Minimum Bazel Version upgraded to 2.2.0 to better support Go rules
• FHIRPath support for skip(), take(), union(), anyTrue(), anyFalse()

Upcoming Features

• Parsing, Printing, and Validating in Python
• FHIRPath validation in Go
• Updated TensorFlow/ML library with support for any version of FHIR
• Java Protos and Libraries published as Maven Jars.

This minor release contains mostly internal refactors, with an eye towards scalability for supporting future versions of FHIR. It will provide a baseline for upcoming feature releases, including support for FHIR R5 and limited support for DSTU2, and FhirProto parsing, printing, and validating libraries in Python and Go.

FHIR Protos

• Core R4 protos upgraded to 4.0.1
• UsCore R4 protos upgraded to 3.1.0
• New ability to generate custom CodeSystem/ValueSet from Profile Configs for defining an Implementation Guide

Language Upgrades

• Upgraded to Python3 throughout
• Upgraded to C++17
• Switched from tensorflow::Status to the newly open-sourced absl::Status

FHIRPath

• Added support for both STU3 and R4
• Added support for DateTime, String, Boolean, Decimal, Integer, and empty ({}) literals
• Added support for ~30 FHIRPath functions
• API improvements for diagnosing unmet FHIRPath constraints
• Support for additional operators
• Support for $this, %context, %resource,%ucum, %sct, and %loinc
• Upgraded to N1 FHIRPath grammar definition

This release adds support for the R4 version of the FHIR spec. We publish R4 versions of the core datatypes, resources, extensions, profiles, code systems and value sets, along with the US-Core and QI-Core profile sets. Most libraries have been updated to work on any version of FHIR, and provide version-specific APIs when that is not possible.

This release also introduces a FHIRPath validation engine to the resource validator, with support for many of the most common expressions in FHIRPath.

Features

• R4 protos for Core FHIR, US-Core and QI-Core profiles
• Library support for R4 parsing, printing, validating, and profiling
• FHIRPath support for boolean logic, arithmetic, date comparisons, and commonly used functions in resource validation (hasValue, exists, not), with complete support to follow
• Better Code enum generation to more accurately differentiate between CodeSystems and ValueSets, and support for more dynamic enum generation for bound Codings and profiled CodeableConcepts
• Better Profile Package management
• Performance improvements

This release expands the tensorflow and profile generation abilities, as illustrated in the examples. The tensorflow example now includes a sequence of steps to generate training samples and build a tensorflow model from Synthea data, and the profile generator now works with FHIR package definitions following the FHIR NPM package spec.

Features

• End-to-end TensorFlow model training example
• Google Cloud Datalab notebook illustrating tensorflow model training
• Proto package generation from NPM-style package.tar.gz files
• Generated US-Core protos for STU3 using NPM package
• Cleaner separation of spec files and generated testdata
• Various bugfixes and performance improvements

This release adds initial support for generating TensorFlow examples. The provided code is an updated version of that used in our publication, Scalable and accurate deep learning with electronic health records. Additional usage examples and sample models will be added over the coming months.

The support for generating custom protos has been much improved in this release. The C++ implementation now comes with a JSON parser and printer, which supports profiles with inlined extensions.

Features

• Code to convert FHIR bundles to TensorFlow SequenceExamples
• Google-defined extensions for labels used in example generation
• C++ JSON parser / printer (in addition to the existing Java JSON parser/printer)
• Generation of DataTypes, Extensions and Resources from StructureDefinitions
• Expanded support for profiles and slicing
• Generated protos for the US Core Implementation Guide
• Updated examples

This release contains the source code used to generate protocol buffer definitions from StructureDefinitions, as well as an example that does just that for the set of StructureDefinitions provided as part of the FHIR spec.

It also contains an example implementation of a client using the new bulk download draft standard, and a few minor bugfixes.

This initial release of the Google FHIR support package provides protos for all resources, along with a Json parser in Java. It comes with an extensive suite of tests to verify compliance with the FHIR standard, and a number of utility classes to simplify usage.

In addition, there's an example showing how to use this representation to query over FHIR data in BigQuery.

More details about this release can be found in our blog post. Additional languages are planned for v0.2.