Getting started

This article describes how to load data, train an MDM model, and use match operation.

Create a client

First you need to have a client for Aidbox mdm Python integration

PUT /Client/mdm-client

secret: secret
grant_types:
  - basic

Create an access policy

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

Prepare Python

Install Python 3, pip, and poetry.

Follow MDM Python integration readme to launch Jupyter Notebook.

Train a model

Follow MDM Python integration tutorial to learn how to train an MDM model

Export a model

Use save_zen_model_edn method to export zen symbol for your model

Add model to Aidbox configuration project

Insert 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}}}

Prepare MDM tables

Call aidbox.mdm/update-mdm-tables RPC method to prepare MDM tables.

POST /rpc

method: aidbox.mdm/update-mdm-tables
params:
  resource-type: Patient

Use match operation

Use $match operation or aidbox.mdm/match RPC method to find similar resources:

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