Getting started
This article describes how to load data, train an MDM model, and use match operation.
First you need to have a client for Aidbox mdm Python integration
PUT /Client/mdm-client
secret: secret
grant_types:
- basic
Create an acccess policy which allows the client to execute raw SQL queries
PUT /AccessPolicy/mdm-policy
engine: allow
link:
- id: mdm-client
resourceType: Client
Use
save_zen_model_edn
method to export zen symbol for your modelInsert exported symbol to your Aidbox configuration project. And enable an MDM service.
Example:
{ns mdmbox
import #{aidbox.mdm
aidbox
zenbox
zen.fhir}
patient-mdm-model model-exported-from-splink
patient-mdm
{:zen/tags #{aidbox/service}
:engine aidbox.mdm/splink-engine
:models #{patient-mdm-model}}
box
{:zen/tags #{aidbox/system}
:services
{:patient-mdm patient-mdm}}}
POST /rpc
method: aidbox.mdm/update-mdm-tables
params:
resource-type: Patient
POST /fhir/Patient/$match
resourceType: Parameters
parameter:
- name: resource
resource:
name:
- given:
- John
family: Smith-Johnson-Williams
birthDate: "2000-01-01"
- name: onlyCertainMatches
valueBoolean: true
Example output:
resourceType: Bundle
type: searchset
entry:
- score: 0.99
resource:
name:
- given:
- John
family: Smith-Johnson-Williams
address:
- city: London
birthDate: "2000-01-01"
resourceType: Patient
id: pt-1
Last modified 4mo ago