...
As learned by the UgandaEMR team at the 2022 UCSF/OpenMRS Mombasa Hackathon:
Step 0: Check if you are breaking any OpenMRS validation rules
Try running the Concept Validation Queries to see if there are any cases of concepts or names that break OpenMRS concept validation rules.
Step1: Get your SQL payload
...
Step 2: Get your JSON payload
Step 3: Create your source on Open Concept Lab
Create a source within the OCL Staging Environments (app.staging.openconceptlab.org). When creating the source, be sure to select the OpenMRS Custom Validation Schema to ensure you don't introduce new concepts that break OpenMRS validation rules.
Step 4: Bulk Import
Use the OCL Bulk Import JSON was bulk imported into the PIH source on OCL using the OCL app's bulk import feature.qa: https://app.qa.openconceptlab.org/
staging: https://app.staging.openconceptlab.org/#/Workflow
| Widget Connector | ||
|---|---|---|
|
...
| CIEL | ocl_omrs Dictionary Conversion Script (converts DB dump into OCL import file) | OCL Bulk Import | OCL TermBrowser | DM WebApp | OCL Subscription Module | OMRS EMR | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| OMRS Nuance | (created b/c Bulk Import doesn't support OMRS SQL - script converts from OMRS model and OMRS DB to OCL model) | Export your existing content (use SQL script here) | |||||||||||||
| Sort Weight | Being created in version of OMRS that doesn't have sort weight on concept answers Burke to work with Andy | Sort Weight on Concept Answers is not consistently populated, and frequently does not match the expected value. So either this isn't getting imported correctly into OCL, isn't getting exported correctly from OCL, or isn't getting imported correctly in the subscription module | Add column for CSV format "sort_weight" and apply to correct mapping | Add support for importing sort_weight
| |||||||||||
| Matching Sources | The ocl_omrs script hardcodes the OpenMRS and OCL versions of these names (e.g., "SNOMED CT" in OpenMRS is "SNOMED-CT" in OCL) | The subscription module creates new concept sources if needed, and has no way to match up against existing concept sources if they are named slightly differently. (No validation step.) For whatever reason, in our existing system we have "SNOMED CT" whereas in OCL this is "SNOMED-CT", so these are all getting associated with a different source. Also all of the expected "PIH" mappings are imported as "PIH-temp" mappings, which makes the diffs hard to use to find actual differences. OCL Module should prompt user: "I'm about to create a new source. Is this what you want? Or, pick one to map to/use." | |||||||||||||
| 2 Yes Concepts | Where are the 2 yes's coming from? To investigate... | ||||||||||||||
| Different Concept Total | PIH experience: Before the import there were 5713 concepts in the dictionary, and afterwards there were 5884 concepts. I will try to dig into this different | ||||||||||||||
| Need codes | Respect new map type called "Code": https://talk.openmrs.org/t/building-conventions-for-concept-ids-in-ocl/34809 | Respect new map type called "Code": https://talk.openmrs.org/t/building-conventions-for-concept-ids-in-ocl/34809 |
...