Skip to content

soneeee22000/GreenLens.dev

BranchesTags

Repository files navigation

GreenLens

.NET Angular Azure License: MIT

AI-powered cloud carbon footprint estimation for Azure infrastructure. Submit your resource usage — get CO2e estimates, per-resource breakdowns, and actionable reduction recommendations powered by Azure OpenAI.

Live Demo | Swagger API

CI Backend Tests E2E Tests Angular Tests .NET Angular C# TypeScript Azure License: MIT

Live Demo | Architecture | API Reference | Tech Stack | Getting Started | Test Coverage | Engineering Decisions

Why GreenLens?

The EU Corporate Sustainability Reporting Directive (CSRD) mandates Scope 3 emissions reporting starting 2026. Cloud infrastructure is a growing portion of enterprise carbon footprint, yet existing tools are delayed, high-level, and lack actionable insights. GreenLens fills that gap with a sub-500ms API that returns precise CO2e estimates and AI-generated reduction strategies.

Architecture

graph TB
    subgraph Frontend["Angular 17 SPA"]
        UI[Angular Material UI]
        Charts[Chart.js Visualizations]
        Forms[Reactive Forms]
    end

    subgraph API["ASP.NET Core 8 Web API"]
        Controllers[REST Controllers]
        Middleware["Auth | Rate Limit | Error Handler"]
        Cache[In-Memory Cache]
    end

    subgraph Core["Domain Layer"]
        Engine[Estimation Engine]
        Models[Domain Models]
        Interfaces[Service Interfaces]
    end

    subgraph Infrastructure["Infrastructure Layer"]
        EF[EF Core + SQLite]
        SearchSvc[Azure AI Search Client]
        OpenAISvc[Azure OpenAI Client]
    end

    subgraph Azure["Azure Cloud Services"]
        AISearch[(Azure AI Search<br/>Emission Factors Index)]
        OpenAI[Azure OpenAI<br/>GPT-4o-mini]
    end

    Frontend -->|HTTP + API Key| API
    API --> Core
    Core --> Infrastructure
    Infrastructure --> Azure

    style Frontend fill:#1976d2,color:#fff
    style API fill:#388e3c,color:#fff
    style Core fill:#f57c00,color:#fff
    style Infrastructure fill:#7b1fa2,color:#fff
    style Azure fill:#0078d4,color:#fff
Loading

Clean Architecture with strict dependency inversion: Api -> Core <- Infrastructure. The domain layer owns interfaces; infrastructure implements them. No Azure SDK leaks into business logic.

Request Flow

sequenceDiagram
    participant Client as DevOps Engineer
    participant API as GreenLens API
    participant Engine as Estimation Engine
    participant Search as Azure AI Search
    participant OpenAI as Azure OpenAI
    participant DB as SQLite

    Client->>API: POST /api/v1/estimates
    API->>Engine: Calculate CO2e
    Engine->>Search: Query emission factors
    Search-->>Engine: Matched factors (kgCO2e/unit)
    Engine->>Engine: Compute per-resource breakdown
    Engine->>DB: Persist estimate
    DB-->>API: Estimate ID
    API-->>Client: 201 Created (total CO2e + breakdown)

    Client->>API: GET /estimates/{id}/recommendations
    API->>OpenAI: Generate reduction strategies
    OpenAI-->>API: Structured recommendations
    API->>API: Cache (1-hour TTL)
    API-->>Client: 200 OK (recommendations)
Loading

Tech Stack

Layer Technology Purpose
Frontend Angular 17, TypeScript (strict), Angular Material, Chart.js SPA with dashboard, forms, and data visualization
Backend ASP.NET Core 8, C# 12 REST API with middleware pipeline
Domain Clean Architecture, CQRS-lite Business logic isolation
AI/Search Azure AI Search, Azure OpenAI (GPT-4o-mini) Semantic emission factor lookup + reduction recommendations
Database EF Core 8 + SQLite (dev) / Azure SQL (prod) Estimate persistence with migrations
Testing xUnit, Moq, Jasmine/Karma, Playwright 88 tests across unit, integration, and E2E
DevOps Docker, GitHub Actions, Swagger/OpenAPI Containerized builds, CI pipeline, auto-generated API docs

API Endpoints

Method Endpoint Purpose Auth
GET /health Health check None
GET /api/v1/regions List 15 Azure regions with grid carbon intensity None
POST /api/v1/estimates Create carbon footprint estimate API Key
GET /api/v1/estimates List estimates (paginated) API Key
GET /api/v1/estimates/{id} Get estimate with per-resource breakdown API Key
GET /api/v1/estimates/{id}/recommendations AI-powered reduction recommendations API Key
GET /api/v1/emission-factors/search Semantic search across emission factors API Key

Example: Create Estimate

curl -X POST https://localhost:5001/api/v1/estimates \
  -H "Content-Type: application/json" \
  -H "X-Api-Key: your-key" \
  -d '{
    "resources": [
      { "resourceType": "Standard_D4s_v3", "region": "westeurope", "quantity": 2, "hours": 720 },
      { "resourceType": "BlobStorage", "region": "westeurope", "quantity": 500, "hours": 720 }
    ]
  }'

Example Response

{
  "data": {
    "estimateId": "a1b2c3d4-e5f6-7890-abcd-ef1234567890",
    "totalCo2eKg": 12.35,
    "createdAt": "2026-03-05T10:00:00Z",
    "breakdown": [
      {
        "resourceType": "Standard_D4s_v3",
        "region": "westeurope",
        "quantity": 2,
        "hours": 720,
        "co2eKg": 8.12,
        "co2ePerUnit": 0.00564,
        "unit": "kgCO2e/hour"
      },
      {
        "resourceType": "BlobStorage",
        "region": "westeurope",
        "quantity": 500,
        "hours": 720,
        "co2eKg": 4.23,
        "co2ePerUnit": 0.00586,
        "unit": "kgCO2e/GB-month"
      }
    ]
  },
  "error": null
}

Project Structure

GreenLens/
├── src/
│   ├── GreenLens.Api/              # Controllers, middleware, DI configuration
│   ├── GreenLens.Core/             # Domain models, interfaces, estimation engine
│   ├── GreenLens.Infrastructure/   # EF Core repos, Azure AI Search, Azure OpenAI
│   ├── GreenLens.Shared/           # DTOs, API contracts, constants
│   └── greenlens-ui/               # Angular 17 SPA
│       ├── src/app/components/     # Dashboard, EstimateForm, EstimateDetail, Search
│       ├── src/app/services/       # Typed HTTP client for all API endpoints
│       └── e2e/                    # Playwright E2E tests (20 tests)
├── tests/
│   ├── GreenLens.Core.Tests/       # 11 unit tests (estimation engine, domain logic)
│   ├── GreenLens.Api.Tests/        # 8 integration tests (endpoints, middleware)
│   └── GreenLens.Infrastructure.Tests/ # 6 tests (repositories, Azure service mocks)
├── tools/
│   └── GreenLens.Seed/             # CLI to seed emission factors into Azure AI Search
├── docs/
│   └── PRD.md                      # Product requirements document
├── .github/workflows/ci.yml        # CI: build, test, format, Docker
├── Dockerfile                      # Multi-stage build, non-root user
├── docker-compose.yml              # Local dev orchestration
└── GreenLens.sln

Test Coverage

pie title 88 Tests Across All Layers
    "Angular Unit Tests" : 43
    "Playwright E2E Tests" : 20
    "Core Domain Tests" : 11
    "API Integration Tests" : 8
    "Infrastructure Tests" : 6
Loading
Layer Framework Count Covers
Domain xUnit + Moq 11 Estimation engine, CO2e calculations, edge cases
API xUnit + WebApplicationFactory 8 Endpoint contracts, auth, error responses
Infrastructure xUnit + Moq 6 Repository CRUD, Azure service integration
Frontend Jasmine + Karma 43 All components, services, form validation
E2E Playwright 20 Full user flows with API mocking

Quick Start

Prerequisites

Backend

git clone git@github.com:soneeee22000/GreenLens.dev.git
cd GreenLens
cp .env.example .env    # Add your Azure credentials
dotnet restore
dotnet ef database update --project src/GreenLens.Infrastructure --startup-project src/GreenLens.Api
dotnet run --project src/GreenLens.Api

API at https://localhost:5001 | Swagger at https://localhost:5001/swagger

Frontend

cd src/greenlens-ui
npm install
ng serve

Angular app at http://localhost:4200

Docker

docker compose up --build

Run All Tests

# Backend (25 tests)
dotnet test

# Frontend unit tests (43 tests)
cd src/greenlens-ui && ng test --watch=false --browsers=ChromeHeadless

# E2E tests (20 tests)
cd src/greenlens-ui && npm run e2e

Key Engineering Decisions

Decision Rationale
Clean Architecture over MVC Domain logic is testable in isolation; Azure services are swappable
Azure AI Search for emission factors Semantic search over 50+ factors vs. hardcoded lookup tables
GPT-4o-mini over GPT-4 10x cheaper, sufficient for structured recommendation generation
SQLite for dev, Azure SQL for prod Zero-config local development; same EF Core migrations for both
API key auth over JWT Simpler for DevOps tool/pipeline integration; no user sessions needed
Angular Material over custom CSS Enterprise-standard components; accessibility built-in
Playwright over Cypress Native multi-browser support; better API mocking via route interception
In-memory cache (1h TTL) for recommendations Avoids redundant OpenAI calls; recommendations don't change frequently

Environment Variables

# Azure AI Search
AZURE_SEARCH_ENDPOINT=https://<name>.search.windows.net
AZURE_SEARCH_API_KEY=<key>
AZURE_SEARCH_INDEX_NAME=emission-factors

# Azure OpenAI
AZURE_OPENAI_ENDPOINT=https://<name>.openai.azure.com/
AZURE_OPENAI_API_KEY=<key>
AZURE_OPENAI_DEPLOYMENT_NAME=gpt-4o-mini

# API Security
API_KEY=<generate-a-strong-key>

# Database
DATABASE_CONNECTION_STRING=Data Source=greenlens.db

License

MIT

Built by Pyae Sone (Seon)

About

AI-powered cloud carbon footprint estimator — CO2e estimates, per-resource breakdowns, and AI reduction recommendations. ASP.NET Core 8 + Angular 17 + Azure AI Search + Azure OpenAI.

greenlens-api.azurewebsites.net

Topics

angular sustainability ai dotnet azure clean-architecture carbon-footprint azure-openai

Resources

Readme
Activity

Stars

0 stars

Watchers

0 watching

Forks

0 forks
Report repository

Releases

No releases published

Packages

 
 
 

Contributors

Languages