Stream: python
Topic: Python breakout Sydney WGM
Ward Weistra (Feb 02 2020 at 10:18):
Cross post: https://chat.fhir.org/#narrow/stream/179207-connectathon-mgmt/topic/Python.20breakout
Eric Haas (Feb 03 2020 at 06:19):
Sydney 2020 HL7 Connecthon Python Breakout Session
- Harold Solbrig presented current state of the Smart FHIR Parser for python.
- fhir-parser untouched
-
fhir templates (fhir-model-generator) modified to utilize Python dataclasses
- hints
- syntactical sucre
-
split the fhirclient library from the models
-
Ilya Beda presented the fhir-py async/sync FHIR client for python3.
-
Next Steps
-
Create pip, pipenv based package based on Harold's fork of the fhir parser and templates
- plan R4+ major versions only
- @Harold Solbrig to provide dig data example for library
- @ Bob Milius to look at making a Jupyter Notebook as a howto generate local library for custom or intermediary versions
-
Ilya - to look at integration of models in FHIR into his sync/async library
-
Future steps
- generate models for profile
- fix primitives to handle Extensions
- FHIRpath? check existing libraries.
Pascal Pfiffner (Feb 03 2020 at 07:22):
Sounds great!
Eric Haas (Feb 10 2020 at 23:30):
Here is a quick demo Jupyter file on Binder using @Harold Solbrig fork of the fhir-model-generator. It shows how to create the models locally and use them to create instances.
Eric Haas (Feb 10 2020 at 23:52):
Doesn't have the shortcuts that Harold has talked about. Hoping we are getting closer to that....
Eric Haas (Feb 12 2020 at 06:13):
Oops that link above is not working try this one instead:
Binder of My Python NoteBooks to share
:notebook: https://mybinder.org/v2/gh/Healthedata1/MyBinder/master
to start the Notebook:
- click on the
pyFHIR_models
folder - click on the notebook file
Howto_python_models.ipynb
( If anybody knows a more direct way to link to the actual Binder Notebook let me know)
P.s. it also shows how to go from excel-->pandas-->FHIR
Ilya Beda (Feb 18 2020 at 09:24):
I have started experimenting with the integration of models from Harold's library.
https://github.com/beda-software/fhir-py-experements/blob/master/fhir_model_experement.py
It seems that by using a simple adapter function you can use any kind of model layer with fhir-py.
Ward Weistra (Feb 18 2020 at 15:40):
Don't understand the code, but that's great news!
Eric Haas (Mar 10 2020 at 03:37):
@Harold Solbrig Could we do this with the fhir_model_generator?
https://github.com/Healthedata1/MyBinder/blob/master/pyFHIR_models/could_we_do_this.ipynb
Last updated: Apr 12 2022 at 19:14 UTC