📊 Purchase & Sales Analytics API

A high-performance .NET 8 Web API for bulk ingestion and analysis of purchase and sales data from Excel/CSV files. Built with Clean Architecture, designed to handle 1M+ rows in under 0.35 seconds.

---

🚀 Features

• Bulk Data Ingestion — Upload Excel (XLSX) or CSV files; data is parsed and inserted via batched bulk operations
• 1M+ Rows in 0.35s — Optimized using EFCore.BulkExtensions with configurable batch sizes (default: 20,000 rows)
• Automatic Product Detection — Products not found in the database are auto-created during sales upload
• Analytics Endpoints — Query profit reports, top/bottom-selling products, and parameterized product search
• Clean Architecture — Domain, Application, Infrastructure, and API layers fully separated
• Global Exception Handling — All errors returned as structured JSON responses
• Swagger Documentation — Interactive API docs available at the root URL

---

🏗️ Architecture

SRC/
├── Purchase&Sales_API/             # Presentation Layer — Controllers, Middleware, Program.cs
├── Purchase&Sales_Core/            # Application Layer — Services, Use Cases, Interfaces
├── Purchase&Sales_Domain/          # Domain Layer — Entities, Enums (no external dependencies)
├── Purchase&Sales_Infrastructure/  # Infrastructure Layer — EF Core, Repositories, BulkExtensions

Why Clean Architecture? Each layer only depends on the one below it. The domain has zero external dependencies — business rules stay isolated from infrastructure concerns like databases and file parsing.

---

🛠️ Tech Stack

Category	Technology
Framework	ASP.NET Core Web API (.NET 8)
ORM	Entity Framework Core
Bulk Operations	EFCore.BulkExtensions
CSV Parsing	CsvHelper
Excel Parsing	EPPlus (OfficeOpenXml)
Database	SQL Server
API Docs	Swagger / Swashbuckle
DI	Built-in .NET DI

---

📚 API Endpoints

Data Ingestion

Method	Endpoint	Description
POST	/api/Purchase/upload	Upload purchase data (XLSX or CSV)
POST	/api/Sales/upload	Upload sales data (XLSX or CSV)

Analytics

Method	Endpoint	Description
GET	/api/Analytics/profit	Product profit report
GET	/api/Analytics/top-sold	Most sold products
GET	/api/Analytics/least-sold	Least sold / deadstock products
GET	/api/Analytics/search	Search products by name or category

---

⚡ Performance Design

The ingestion pipeline is built around three principles:

1. Streaming parsing — Files are read row-by-row, never loaded fully into memory
2. Batch processing — Rows are grouped into batches of 20,000 before hitting the database
3. Bulk insert — EFCore.BulkExtensions bypasses EF's change tracker and writes directly via SqlBulkCopy

This combination is what achieves the 1M rows in 0.35 seconds benchmark.

---

📦 Getting Started

Prerequisites

• .NET 8 SDK
• SQL Server (local or remote)

Setup

1. Clone the repository

   git clone https://github.com/AhmedMTwab/Purchase-Sales-Analysis-Task.git
   cd Purchase-Sales-Analysis-Task

2. Configure the database connection in appsettings.json

   {
     "ConnectionStrings": {
       "DefaultConnection": "Server=.;Database=PurchaseSalesDb;Trusted_Connection=True;"
     }
   }

   📌 Database Connection Note: This application is connected to a deployed database, so you don't need to change the connection string. However, if the deployed database fails or you want to use your own database, you can update the connection string above.

3. Apply migrations

   dotnet ef database update --project SRC/Purchase&Sales_Infrastructure

4. Run the application

   dotnet run --project SRC/Purchase&Sales_API

5. Open Swagger UI at https://localhost:5001 (or the port shown in your console)

---

🗄️ Database Diagram

---

📝 Notes

• Large files are processed in configurable batches — adjust BatchSize in appsettings.json based on your server memory
• Auto-created products during sales upload are flagged for review via the product search endpoint
• All unhandled exceptions are caught globally and returned as structured JSON with appropriate HTTP status codes

---

👤 Author

Ahmed Mohamed Eltwab