Skip to content

AGENTS.md

Latest commit

 

History

History
81 lines (57 loc) · 2.62 KB

AGENTS.md

File metadata and controls

81 lines (57 loc) · 2.62 KB

AGENTS.md

This file provides guidance to AI coding agents such as Claude Code (claude.ai/code) and Gemini Code Assist when working with code in this repository.

Best Practices

When working with this codebase, always follow modern .NET and C# best practices:

  • Use latest C# language features (records, pattern matching, nullable reference types, file-scoped namespaces)
  • Follow .NET naming conventions and coding standards
  • Leverage modern async/await patterns where applicable
  • Use System.Text.Json for JSON serialization
  • Prefer Span<T> and Memory<T> for performance-critical code
  • Use collection expressions and LINQ for data manipulation
  • Apply proper error handling with exceptions and Result patterns
  • Write clean, readable code with meaningful variable names and comprehensible tests
  • Use nullable reference types and enable all relevant compiler warnings

Commands

Build

dotnet build

Test

dotnet test --no-restore

Run a single test

dotnet test --filter "FullyQualifiedName~TwoSumTest"

Run console application

cd ./LeetCode
dotnet run

# Run specific commands
dotnet run benchmark LRUCache --csharp
dotnet run info MergeTwoLists
dotnet run list

Architecture

This is a LeetCode problem solutions repository with benchmarking capabilities, structured as follows:

Three main projects:

  1. LeetCode.CSharp - C# implementations of LeetCode problems with xUnit tests and BenchmarkDotNet benchmarks
  2. LeetCode.FSharp - F# implementations of selected problems
  3. LeetCode - Console application for running benchmarks and viewing problem information

Key architectural patterns:

Problem Organization:

All C# problems are implemented as static methods in the Problem partial class, decorated with [LeetCode] attributes containing metadata (description, difficulty, category, NeetCode video link). Each problem includes inline xUnit tests using [Fact] attributes.

Benchmark Structure:

Each problem has a corresponding benchmark class in the Benchmarks folder that inherits from a base Benchmark class. Benchmarks use BenchmarkDotNet with GlobalSetup/GlobalCleanup for test data preparation.

Console Application:

Built with Spectre.Console, provides an interactive menu system and CLI commands (app, benchmark, info, list, workflow) for running benchmarks and viewing problem information.

Testing:

Uses xUnit with Shouldly for test assertions. Tests are co-located with problem implementations for easy access.

Package Management:

Uses Directory.Packages.props for centralized NuGet package version management across all projects.