Find out what resources were exported during the $export operation

This guide explains how to find out what resources certain Bulk client had access to

$export operation is used to export huge amounts of patient data directly to S3 buckets. Sometimes it is important to clarify what resources certain Bulk client had access to.

In the guide we analyze the behavior of the inferno-my-clinic-bulk-client client

When the client used $export

To clarify when our client exported the data:

1. Open the kibana UI

2. Enter the search request like w_cid:inferno-my-clinic-bulk-client and w_m:get and w_url like /bulk-api/Group//$export

   • w_cid is the id of out client

3. Define the dates range

4. Press the Refresh button

When the request list appears, collect if of the exported groups and timestamps.

1. Extract group id from the w_url property /tenant/my-clinic/bulk-api/Group/test-group-1/$export -> test-group-1

2. Extract timestamp from the ts property -> Dec 5, 2022 @ 16:34:30.417

Get list of patients

To get list of affected by the $export patient records:

1. Open the REST Console in Aidbox

2. Perform the GET /Group/test-group-1/_history request for each group collected on the previous step

3. Search for the version being active during the export (ts property). The matching version has lastUpdated less than ts. If there are several versions, the closes to ts should be taken

4. List of patient are held in the members section

 member:
 - entity:
     id: test-pt-1
     resourceType: Patient

Get list of the resource were exported

List of resource types client had access to

List of resources types are defined with the scope attribute.

1. Open the REST Console in Aidbox

2. Perform the GET /Client/inferno-my-clinic-bulk-client/_history request

3. Search for the version being active during the export (ts property). The matching version has lastUpdated less than ts. If there are several versions, the closes to ts should be taken

4. List of resources the client had access to is defined by the scope property

5. Convert scopes to the resource types system/Observation.read -> Observation resource type

See how system/*.read expands

List the id of resources

This should be done for every patient and resource type

To build the list of patient compartment resources:

1. Open the REST Console in Aidbox

2. Clarify how to search for the patient's resources (see patient compartment)

3. Get the list of the resources. For example, GET /fhir/Observation?subject=test-pt-1&_elements=id

4. Collect all the ids related to the patient and a resource type.

The output could be a table like this.

patient_id	resource_type	resource_id
test-pt-1	Observation	test-obs-1
test-pt-1	Observation	test-obs-2
	Location	loc-1
test-pt-1	Patient	test-pt-1

Location here is the example of the resource not in the compartment so the patient_id is empty

Get the exported resources details

This should be done for every row on the table above

To get the exported resources you should repeat those steps changing the resource type

1. Open the REST Console in Aidbox

2. Perform the GET /fhir/{resource_type}/{resource_id}/_history request. For example, GET /fhir/Observation/test-obs-1/_history

3. Search for the version being active during the export (ts property). The matching version has lastUpdated less than ts. If there are several versions, the closes to ts should be taken