Skip to content

BillMillerCoding/WarrantyTrackerWebApp

BranchesTags

Repository files navigation

WarranTea — Warranties Made Easy

A full-stack warranty tracking application with AI-powered features including OCR receipt scanning, semantic product search, and an AI warranty assistant. Built with ASP.NET Core 10 and Nuxt 4, deployed on Azure.

Live Demo:

Architecture

┌─────────────────────┐       ┌─────────────────────┐
│   Nuxt 4 Frontend   │──────▶│  ASP.NET Core API   │
│  Azure Static Web   │ HTTPS │  Azure App Service   │
│       Apps           │◀──────│                      │
└─────────────────────┘       └──────────┬───────────┘
                                         │
                    ┌────────────────────┼────────────────────┐
                    │                    │                    │
            ┌───────▼──────┐   ┌────────▼───────┐   ┌───────▼──────┐
            │  Azure SQL   │   │  Azure Blob    │   │  Azure AI    │
            │  Database    │   │  Storage       │   │  Services    │
            └──────────────┘   └────────────────┘   └──────────────┘

Azure Services

Service Purpose
Azure App Service Hosts the ASP.NET Core API
Azure Static Web Apps Hosts the Nuxt frontend
Azure SQL Database Primary relational database (EF Core + Identity)
Azure Blob Storage Receipt images, warranty documents, and OCR extracted text files
Azure AI Search Hybrid semantic + vector search over product warranties (HNSW, cosine similarity, 1024-dim vectors)
Azure AI Vision Multimodal embeddings for text and image vectors, plus OCR text extraction from receipt images
Azure OpenAI AI chat assistant for warranty-related Q&A (GPT-based)

Features

User Warranty Management

  • Create, view, edit, and delete personal warranties
  • Upload receipt images with automatic OCR text extraction
  • Upload warranty documents
  • Filter warranties by search query, brand, and status
  • Automatic status computation (active, expiring soon, expired)

AI-Powered OCR

  • Upload a receipt photo and the system extracts warranty fields (product name, brand, purchase date, expiration date, coverage type) using Azure AI Vision OCR and regex parsing

Semantic Product Search

  • Hybrid text + vector search across 71 pre-loaded product warranties
  • Image-based search — upload a product photo to find matching warranties
  • Brand and category filtering with semantic ranking

AI Warranty Assistant

  • Chat with an AI assistant about warranty questions
  • Conversation history support
  • Powered by Azure OpenAI

Admin Product Warranty Management

  • Full CRUD for the product warranty catalog
  • Re-index all product warranties into Azure AI Search

Authentication

  • Cookie-based authentication via ASP.NET Identity
  • Role-based authorization (User, Admin)
  • Cross-domain cookie support (SameSite=None, Secure, HttpOnly)

Tech Stack

Backend

Technology Version
ASP.NET Core 10.0
Entity Framework Core 10.0.4
ASP.NET Identity 10.0.4
Azure.Search.Documents 11.7.0
Azure.Storage.Blobs 12.27.0
C# / .NET 10

Frontend

Technology Version
Nuxt 4.3.1
Vue 3.5.30
Vuetify 3.12.2
TypeScript —
Vitest 4.1.0

API Endpoints

Auth — api/auth

Method Route Auth Description
POST /api/auth/register Public Register a new user
POST /api/auth/login Public Login (sets auth cookie)
POST /api/auth/logout Public Logout (clears cookie)
GET /api/auth/me Authorized Get current user info

Warranties — api/warranties

Method Route Auth Description
GET /api/warranties Authorized List user's warranties (filterable)
GET /api/warranties/{id} Authorized Get warranty by ID
POST /api/warranties Authorized Create warranty
PUT /api/warranties/{id} Authorized Update warranty
DELETE /api/warranties/{id} Authorized Delete warranty
POST /api/warranties/{id}/upload-receipt Authorized Upload receipt (auto-OCR)
GET /api/warranties/{id}/receipt Authorized Download receipt file
GET /api/warranties/{id}/extracted-text Authorized Download OCR text
POST /api/warranties/{id}/upload-doc Authorized Upload warranty document

Products — api/products

Method Route Auth Description
GET /api/products Public List all products
GET /api/products/{id} Public Get product with warranties

Product Search — api/products/search

Method Route Auth Description
GET /api/products/search Public Hybrid text + vector search
POST /api/products/search/image Public Search by image upload

OCR — api/ocr

Method Route Auth Description
POST /api/ocr/parse-warranty Authorized Extract warranty fields from image

AI — api/ai

Method Route Auth Description
POST /api/ai/warranty-query Authorized Chat with AI warranty assistant

Admin — api/admin/product-warranties

Method Route Auth Description
GET /api/admin/product-warranties Admin List all product warranties
GET /api/admin/product-warranties/{id} Admin Get product warranty
POST /api/admin/product-warranties Admin Create product warranty
PUT /api/admin/product-warranties/{id} Admin Update product warranty
DELETE /api/admin/product-warranties/{id} Admin Delete product warranty
POST /api/admin/product-warranties/reindex Admin Re-index search index

Health — api/health

Method Route Auth Description
GET /api/health Public Database connection status and record counts

Frontend Pages

Route Description
/ Landing page
/login Login form
/register Registration form
/dashboard User warranty dashboard with stats and filtering
/warranties/add Add warranty with OCR receipt scanning
/warranties/search Search personal warranties
/warranties/[id] Warranty detail view
/products/search Semantic product warranty search (text + image)
/ai AI warranty assistant chat
/admin/product-warranties Admin product warranty management

Data Model

Product

20 seeded products across electronics, appliances, and gaming (e.g., MacBook Pro 16", Galaxy S25 Ultra, Dyson V15 Detect, LG C4 OLED TV, PlayStation 5 Pro).

Product Warranty

71 seeded warranty options (3–5 per product) covering Manufacturer, Extended, Accidental, Theft & Loss, International, and specialty coverage types.

User Warranty

User-created warranties linked to products, with receipt/document storage and automatic status tracking.

Seeded Users

Email Role Password
admin@warrantea.com Admin Admin1!
alice@warrantea.com User Alice1!
bob@warrantea.com User Bobob1!
carol@warrantea.com User Carol1!

Testing

93 total unit tests (55 backend + 38 frontend)

Backend — xUnit + Moq + EF Core InMemory

Test File Tests Coverage
WarrantyServiceTests 22 CRUD, status computation, filtering, blob operations
ProductWarrantyServiceTests 13 CRUD, reindex, search integration
AiServiceTests 8 AI calls, error handling, timeouts
OcrServiceTests 7 OCR parsing, regex extraction
ProductServiceTests 5 Product retrieval

Frontend — Vitest + happy-dom + Vue Test Utils

Test File Tests Coverage
types.test.ts 8 TypeScript interface validation
warrantyService.test.ts 6 API calls, FormData, OCR
adminService.test.ts 6 CRUD, reindex
productService.test.ts 6 Search, duration formatting
useAuth.test.ts 6 Cookie auth, roles, login
useAiChat.test.ts 5 Chat history, error handling
auth.test.ts 1 Auth headers

CI/CD

Two GitHub Actions workflows deploy on push to main:

  • API Workflow — Builds .NET solution, runs backend tests, publishes WarranTeaApi.csproj, deploys to Azure App Service via OIDC
  • Frontend Workflow — Installs dependencies, runs frontend tests, generates static site with npm run generate, deploys to Azure Static Web Apps

Local Development

Prerequisites

  • .NET 10 SDK
  • Node.js 18+
  • SQL Server LocalDB (or update the connection string)

Backend

cd WarranTeaApi
# Copy appsettings.template.json to appsettings.json and fill in your Azure keys
dotnet run

The API starts at http://localhost:5000.

Frontend

cd WarranTeaFrontEnd
npm install
npm run dev

The frontend starts at http://localhost:3000 with API calls proxied to localhost:5000.

Run Tests

# Backend
dotnet test WarranTeaApi.Tests

# Frontend
cd WarranTeaFrontEnd
npx vitest run

Configuration

The API reads configuration from appsettings.json locally or Azure App Service environment variables in production:

Key Description
ConnectionStrings:DefaultConnection SQL Server connection string
AzureAISearch:Endpoint Azure AI Search service endpoint
AzureAISearch:AdminApiKey Azure AI Search admin key
AzureAISearch:IndexName Search index name (default: warranties)
AzureAIVision:Endpoint Azure AI Vision endpoint
AzureAIVision:ApiKey Azure AI Vision API key
AzureBlob:ConnectionString Azure Blob Storage connection string
AzureBlob:ContainerName Blob container name (default: files)
AzureOpenAI:Endpoint Azure OpenAI endpoint (including deployment path)
AzureOpenAI:ApiKey Azure OpenAI API key

In Azure App Service, use double underscores as section separators (e.g., AzureAISearch__Endpoint).

  • Upload warranty image

  • OCR + AI extraction of structured data:

    • Product name
    • Brand
    • Purchase date
    • Warranty duration
    • Serial number
    • Store

  • Convert image to searchable PDF

  • Store structured warranty in database

  • List/search warranties per user

  • Manual product warranty lookup

Future Enhancements

  • QR code receipt scanning
  • Email receipt parsing
  • Manufacturer API integrations
  • Warranty expiration notifications
  • Family/shared accounts
  • Chrome extension for online purchases
  • Mobile app version

3. Database Design

User

User
-----
Id (GUID)
ExternalAuthId (string)
Email
CreatedAt

Product

Product
-------
Id (GUID)
Brand
ModelName
Category
UPC (nullable)
CreatedAt

Warranty

Warranty
--------
Id (GUID)
UserId (FK)
ProductId (FK)
PurchaseDate
ExpirationDate
WarrantyDurationMonths
StoreName
SerialNumber
PdfBlobUrl
OriginalImageBlobUrl
AIExtractConfidenceScore
CreatedAt

WarrantyDocument (Optional, Flexible Design)

WarrantyDocument
----------------
Id
WarrantyId (FK)
BlobUrl
DocumentType (OriginalImage, PDF, Receipt)
UploadedAt

ProductWarrantyInfo (Lookup Cache)

ProductWarrantyInfo
-------------------
Id
ProductId (FK)
WarrantyDurationMonths
CoverageDescription
SourceUrl
LastVerifiedAt

4. AI Integration (Azure Foundry)

Use Case 1: OCR + Structured Extraction

Pipeline:

  1. User uploads image
  2. Backend sends image to Azure Document Intelligence / Vision OCR
  3. Extract text
  4. Send text to LLM via Azure Foundry
  5. Prompt for structured JSON output
  6. Deserialize into C# model
  7. Store in database

Important Design Considerations:

  • Require JSON-only output
  • Store AI confidence score
  • Allow manual user edits
  • Display extracted raw text for transparency

Use Case 2: Product Warranty Lookup

Workflow:

  1. User searches product

  2. Check local cache (ProductWarrantyInfo)

  3. If not found:

    • Call Azure AI agent
    • Retrieve warranty terms
    • Return structured JSON

  4. Store in cache table

Must include disclaimer:

Warranty information is AI-generated and may not be legally accurate.

5. Authentication Design

Recommended Approach:

  • Use Azure Static Web Apps built-in auth
  • Entra ID / GitHub / Google provider
  • Pass JWT to backend
  • Backend validates token

User table stores only:

  • ExternalAuthId
  • Email

No password storage required.

6. Blob Storage Organization

Suggested structure:

/users/{userId}/warranties/{warrantyId}/original.jpg
/users/{userId}/warranties/{warrantyId}/warranty.pdf

Store only Blob URLs in database.

7. Key Design Decisions

Multiple Warranties per Product

Allow multiple warranties per product per user.

AI Failure Handling

  • Editable warranty form
  • Store AI confidence
  • Show extracted text

Normalization

Normalize products instead of storing product names directly in Warranty.

8. Scalability Considerations

Infrastructure Scaling

  • Blob Storage for large files
  • SQL for structured data
  • Stateless API layer
  • Horizontal scaling via App Service or Containers

Background Processing (Future)

  • Azure Service Bus
  • Azure Queue Storage
  • Background worker for:
    • PDF generation
    • AI extraction
    • Notifications

Caching Strategy

  • Cache AI lookup results in ProductWarrantyInfo
  • Reduce repeated AI calls
  • Add expiration/refresh logic

Event-Driven Expansion

  • Emit events on warranty creation
  • Trigger:
    • Email reminders
    • Analytics updates
    • AI summary generation

9. Risks & Constraints

  • Warranty data inconsistency
  • AI extraction inaccuracies
  • Manufacturer site scraping reliability
  • Legal accuracy concerns

Mitigations:

  • Manual editing
  • Transparency
  • Confidence scoring
  • Disclaimers

10. Standout Features

  • Warranty expiration dashboard
  • Email notifications (SendGrid)
  • Category filtering
  • AI plain-English warranty summary
  • Multi-document uploads
  • Receipt + warranty matching

11. Overall Architectural Strengths

  • Clean separation of concerns
  • Cloud-native storage model
  • AI integration is purposeful
  • Scalable file handling
  • Secure external authentication
  • Resume-worthy full-stack + AI project

About

No description, website, or topics provided.

Resources

Readme
Activity

Stars

0 stars

Watchers

0 watching

Forks

0 forks
Report repository

Releases

No releases published

Packages

 
 
 

Contributors

Languages