A WPF desktop application for small veterinary laboratories designed to streamline the process of managing and exporting blood test results. Originally developed as a prototype for a real-world lab technician working with animals and humans, this project evolved into a showcase demonstrating clean architecture principles and high code maintainability.
- History-aware navigation by lab order date and number
- Manual pairing of results and patients based on timestamp and test order
- Document templating engine: supports Word and Excel with dynamic token substitution
- Autocomplete for breeds, categories, and animal types
- Clinic-specific template filter based on selected clinic
- Dark/light themes with single-button toggle
- Persistent data with SQLite for long-term storage
- Built-in HL7 device emulator to simulate analyzer data input
- Export to disk with file sanitization and flexible naming conventions
Bidirectional selection between report data and lab results in action. Incoming observation results are processed, then saved and exported
While observing a biochemistry specialist manually transcribe blood analyzer output into Word and Excel documents, it became apparent that an enormous amount of time was wasted on repetitive, low-skill work. Inspired by this inefficiency, I initiated development of a desktop tool that could:
- Receive HL7 messages from blood analyzers
- Pair results with patient data via a simple UI
- Automatically generate print-ready documents
- Adapt to different lab clients and templates with minimal configuration
The initial prototype reduced report preparation time from 2 hours to under 1, eliminating hundreds of keystrokes and potential transcription errors.
This project applies Clean Architecture and DDD-inspired patterns in the context of a desktop application:
- Clean Architecture with distinct layers:
Core: entities, value objects, notification service and validation error codesUseCases: use cases, MediatR commands/queriesInfrastructure: template rendering, HL7 device integration, file handlingDesktop: WPF UI, localization resources, dependency injection setup
- DDD patterns applied:
- Entities enforce business rules and maintain invariants
- Value objects encapsulate validation and consistent state
- Repositories abstract persistence
- Specifications express filtering logic in a reusable way
- .NET 9 & WPF with IHost-based startup (
Microsoft.Extensions.Hosting)- Implements a ViewModel-first pattern with navigation
- Responsive layout and basic theming
- Localization-ready UI (easily extendable via
.resxresources) - Expandable HL7 MLLP device support: adding a new device involves only domain mapping
An HL7 over TCP emulator is bundled, allowing demonstration of analyzer integration without physical devices.
/
├── docs/ # Documentation and token reference
│ ├─screenshots/ # Demo screenshots and gif
│ └─ TokenReference.md
│
├── simulators/ # Device simulators
├── tools/ # CLI tools and utilities
├── src/ # Main application sources
│ ├─ Core/
│ ├─ UseCases/
│ ├─ Infrastructure/
│ └─ Desktop/ # WPF app with resources
│
├── templates/ # Ready-to-use document templates and example data
├── tests/ # xUnit tests for domain and infrastructure
└── README.md
- Animal settings defining animal type, breed, and categories is stored in
animalsettings.json. - A fully localized example list is available in
templates/animalsettings.json. - Templates for export documents (
.docx,.xlsx) are located intemplates/. - Labs can define their own templates and connect them via config.
- Launch the app
- Connect analyzer or emulator (URIT-5160 format supported)
- Register patient data
- Match incoming test results via timestamp-based UI
- Export document with correct template preselected
- Print or archive generated files
main: clean showcase version with ~20 feature-focused commitshistory: original development history with partial preservation of commit evolution
- Add patient search and result history
- Add new analyzer integration via HL7
- Add support for English language UI
- Enable multi-user profiles with shared SQLite instance
MIT