RSBC DMER Optimization & Intake Automation

RSBC DMER optimization pilot for Ministry of AG/PSSG.

This repository is an enterprise monorepo for the Driver Medical Examination Report (DMER) Intake Automation solution. It implements an AI-enabled pipeline that retrieves DMER documents from the Mercury (Dynamics) system, performs automated document extraction (Azure Document Intelligence), normalization (Azure OpenAI, hosted in a separate AI Hub subscription and consumed as an external endpoint — see docs/architecture/repository-design.md §10), rule-based evaluation (GoRules/Zen), and writes decisions/audit data back so Mercury remains the system of record.

Start here

I want to...	Go to
Understand the solution architecture	docs/architecture/
Understand a specific service's responsibilities	docs/services/
Understand queue/message contracts	docs/contracts/queues/
Set up a local dev environment	docs/development/local-development.md
Deploy infrastructure	docs/deployment/deployment-guide.md
Operate / troubleshoot production	docs/operations/
Follow coding standards	docs/development/coding-standards.md, docs/standards/

Repository layout

services/          Independently deployable services (Functions & Container Apps)
libs/               Shared Python libraries used across services
infrastructure/     Bicep IaC (modules + orchestration templates)
deployment/         Per-environment parameter files (dev/test/prod)
database/           PostgreSQL schema, migrations, seed data
monitoring/         Azure Monitor workbooks, alert definitions, dashboards
docs/               Architecture, service, contract, deployment, and ops documentation
scripts/            Developer, CI, and database utility scripts
.github/            GitHub Actions workflows and PR/issue templates

Services

Service	Azure compute	Responsibility
intake-processor	Azure Functions	Batch poll + webhook intake from Mercury, publish to raw-dmer-queue
di-processor	Azure Container Apps	OCR via Document Intelligence, hashing, publish to extracted-dmer-queue
workflow-orchestrator	Durable Functions	End-to-end orchestration: dedupe, case mgmt, normalization, rules, Mercury update
normalizer-service	Azure Container Apps	Structured extraction & evidence validation via an externally-hosted Azure OpenAI model
rule-engine	Azure Functions	GoRules/Zen rule evaluation against rules.json
post-processing	Azure Functions	Audit logging, metrics, notifications, final status
audit-service	Azure Container Apps	Read-side audit/query API for Mercury dashboard & operations

See docs/services/ for details on each.

Status

Infrastructure and service skeletons only — see individual service READMEs for implementation status.