Architecture

Project Structure 

Analogi/
├── Analogi.Core/           Pure .NET 10 library (no UI dependencies)
│   ├── Interfaces/         ILanguageProfile, IPipelineStep, IScanner
│   ├── Models/             CodeFile, PipelineContext, ScanResult, Submission
│   ├── Algorithm/          CosineSimilarity (static), Winnowing (MOSS-style)
│   ├── Languages/          Cpp, Python, Java, CSharp, JavaScript profiles
│   ├── Extractors/         Code, Comment, Structure
│   ├── PreProcessors/      CaseFold, WhitespaceNormalize, StringLiteral, Identifier
│   ├── Analyzers/          Cosine, Winnowing, Structure, Import, Comment, Metrics
│   └── Pipeline/           AnalysisEngine (orchestrator)
│
├── Analogi.App/            AvaloniaUI desktop app (FluentAvalonia theme)
│   ├── Views/              MainWindow, Pages (Scan, Results, Compare, About)
│   └── ViewModels/         MVVM with CommunityToolkit.Mvvm
│
├── Analogi.CLI/            Console app for batch processing
├── Analogi.Tests/          xUnit test suite
└── docs/                   This documentation (Jekyll + just-the-docs)

Design Principles

  1. Core is pure — Analogi.Core has zero UI or platform dependencies. It targets net10.0 and can be used by any .NET application (CLI, web API, desktop app).

  2. Pipeline is pluggable — Every step (extractor, preprocessor, analyzer) implements IPipelineStep. You can add, remove, or reorder steps without changing the engine.

  3. Languages are data — Adding a language means providing regex patterns, not modifying the analysis logic.

  4. Comparison is O(n²/2) — We only compare pairs (i, j) where i < j, halving the work since similarity is symmetric.

  5. Async-first — All scanning is async with IProgress<T> reporting and CancellationToken support.

Data Flow 

Folder Path
    │
    ▼
┌─────────────────┐
│  AnalysisEngine  │
│                  │
│  1. ScanFiles()  │─── Discovers files, groups by language
│  2. For each     │
│     pair (i,j):  │
│                  │
│  ┌────────────┐  │
│  │ Extractors │  │─── CodeExtractor, CommentExtractor, StructureExtractor
│  └─────┬──────┘  │
│        ▼         │
│  ┌────────────┐  │
│  │ PreProcess │  │─── CaseFold → Whitespace → StringLiteral → Identifier
│  └─────┬──────┘  │
│        ▼         │
│  ┌────────────┐  │
│  │ Analyzers  │  │─── Each adds SimilarityReason to context if score > threshold
│  └─────┬──────┘  │
│        ▼         │
│  FilePairResult   │─── Weighted average of all reasons
│                  │
└────────┬─────────┘
         ▼
    ScanResult
    (sorted pairs)

Technology Stack

Component Technology
Core library .NET 10, C# 14
Desktop UI AvaloniaUI 12, FluentAvalonia 2.2
Code viewer AvaloniaEdit 12
MVVM CommunityToolkit.Mvvm 8.2
Testing xUnit
Documentation Jekyll + just-the-docs