fhirpath.py

FHIRPath implementation in Python

Parser was generated with antlr4

Getting started

Install

pip install fhirpathpy

Usage

from fhirpathpy import evaluate

patient = {
  "resourceType": "Patient",
  "id": "example",
  "name": [
    {
      "use": "official",
      "given": [
        "Peter",
        "James"
      ],
      "family": "Chalmers"
    },
    {
      "use": "usual",
      "given": [
        "Jim"
      ]
    },
    {
      "use": "maiden",
      "given": [
        "Peter",
        "James"
      ],
      "family": "Windsor",
      "period": {
        "end": "2002"
      }
    }
  ]
}

# Evaluating FHIRPath
result = evaluate(patient, "Patient.name.where(use='usual').given.first()", {})
# result: `['Jim']`

evaluate

Evaluates the "path" FHIRPath expression on the given resource, using data from "context" for variables mentioned in the "path" expression.

Parameters

resource (dict|list): FHIR resource, bundle as js object or array of resources This resource will be modified by this function to add type information.

path (string): fhirpath expression, sample 'Patient.name.given'

context (dict): a hash of variable name/value pairs.

model (dict): The "model" data object specific to a domain, e.g. R4. See Using data models documentation below.

options (dict) - Custom options.

options.userInvocationTable - a user invocation table used to replace any existing functions or define new ones. See User-defined functions documentation below.

compile

Returns a function that takes a resource and an optional context hash (see "evaluate"), and returns the result of evaluating the given FHIRPath expression on that resource. The advantage of this function over "evaluate" is that if you have multiple resources, the given FHIRPath expression will only be parsed once

Parameters

path (string) - the FHIRPath expression to be parsed.

model (dict) - The "model" data object specific to a domain, e.g. R4. See Using data models documentation below.

options (dict) - Custom options

options.userInvocationTable - a user invocation table used to replace any existing functions or define new ones. See User-defined functions documentation below.

Using data models

The fhirpathpy library comes with pre-defined data models for FHIR versions DSTU2, STU3, R4, and R5. These models can be used within your project.

Example:

from fhirpathpy.models import models


r4_model = models["r4"]

patient = {
  "resourceType": "Patient",
  "deceasedBoolean": false,
}

result = evaluate(patient, "Patient.deceased", {}, r4_model)

# result: `[False]`

User-defined functions

The FHIRPath specification includes a set of built-in functions. However, if you need to extend the functionality by adding custom logic, you can define your own functions by providing a table of user-defined functions.

Example:

user_invocation_table = {
    "pow": {
        "fn": lambda inputs, exp=2: [i**exp for i in inputs],
        "arity": {0: [], 1: ["Integer"]},
    }
}

result = evaluate(
    {"a": [5, 6, 7]},
    "a.pow()",
    options={"userInvocationTable": user_invocation_table},
)

# result: `[25, 36, 49]`

It works similarly to fhirpath.js

Development

To activate git pre-commit hook: autohooks activate

To run tests: pytest