Thank you @sunbiz for the clarification. I got your point. So we need to find out a way to map FHIR to Librehealth JSON and vice versa. If we need to keep a mapping, we might need to use XSLT transformations?
XSLT is a good optionā¦ or you might need JAXB bindings, if you import the HAPI and REST Webservices module jar (remember to use the omod like a JAR, so thats it doesnt need you to write an OpenMRS module). You might also be able to use JOLT, like I said in my earlier post.
3 Likes
@sunbiz I looked at both XSLT and JOLT. Both a feasible options. However I have small concern on the Librehealth JSON to FHIR JSON mappings.
In librehelath if I take patient, the address location is specified through a resource URI such as https://radiology.librehealth.io/lh-toolkit/ws/rest/v1/patient/address/1bb9664c-da14-4db6-aea6-018046c8841c. This is same for the name as well. Hence we need to fetch address, name again with the requested url to fill the FHIR JSON. So we will need to parse the same template multiple time? or should we construct single JSON representation and do a transformation if we goes with template approach?
Also I think we need to have a HAPI library to do a offline validation.
1 Like
I think youāll have to fetch it multiple times. To avoid this, get a full representation by passing the parameter ?v=full after a resource.
1 Like
Hi @sunbiz , I am Amitabh new to the community. Please help me to get started on this project. Are there any tasks that I need to accomplish or a specific aspect of this project as pre-requisite to GSoC application submission. Please point out to those. Iāve introduced myself on the main thread and a community-member directed me to this project. I am setting up the dev-environment and getting started.
Github Profile: https://www.github.com/amitabh27 Linkedin Profile : https://www.linkedin.com/in/tiwariamitabh/
Thanks and Regards, Amitabh Tiwari
@amitabh27, if you are interested in the lh-toolkit or lh-radiology projects, then I recommend that you complete the following prerequisite tasks:
- https://gitlab.com/librehealth/lh-toolkit/issues/34
- https://gitlab.com/librehealth/lh-radiology/issues/42
The REST JSON to FHIR JSON mapping is a good starter project, but its only to give you an idea about comparing the two data models.
Yes v=full seems to be the option 
@prashadi do you have a proposal for us to review for the Spring data, or another toolkit/radiology project that you will be working on over the summer? We are suggesting that you do the drafts in the GSoC application portal only.
1 Like
@sunbiz @judywawira thank you for accepting me to GSoC. Iām looking forward to do my best to achieve success in the project. What should be our next steps? Should I focus on the spring data project or FHIR analytics related project? So I can prepare in advance during the community bonding period and collect necessary information.
1 Like
Hi @prashadi, I will be mentoring you along with @judywawira. I am excited to work with you in the summer. I have been involved in OpenMRS and GSoC for many years and have worked as a Java developer for 9 years.
My suggestion is that you work on the FHIR analytics project. I assume there will be four parts in the project as below:
- Integration of Spark with spring boot application.
- Deployment of FHIR Bundle on a Spark-supported database.
- Spark configuration for FHIR Bundle.
- User interface for query builder.
We could discuss these in detail. You should start a new thread as suggested in the welcome to GSoC thread. @sunbiz suggested that you and @yashdsaraf work closely in a common reactive spring boot 2.x app that works with Spring 5.
Hi @namratanehete, will I be working in the same thread as @prashadi or should I create a new one? Also what should be my next step?
Please make a new one.
1 Like
@namratanehete Thank you again for giving this opportunity to work on this project. It will be a great experience and exciting journey for me to work under the mentorship of @namratanehete and @judywawira.
According to the @sunbiz, I will need to work with @yashdsaraf to integrate the spark with the spring boot application that will be created for spring data integration. So I will use Bensen library and build the FHIR analytics capabilities and build a user interface to query data. I hope my idea is correct.
@sunbiz @namratanehete Also since weāre supposed to work on a common spring app, should I create a gitlab repository holding a spring 2.x reactive app? Perhaps @prashadi and I can work on separate branches in it.
@yashdsaraf yes I think you should create a new gitlab repository and work on different branches.
1 Like
Yes @prashadi. After integration of Spark with spring boot application, your main focus will be on dynamic query builder and analytics capabilities.