TextTV Viewer

A C# console application for displaying Text TV pages from the texttv.nu API with proper styling and formatting. Now supports converting web pages to TextTV format using OpenAI's GPT-4.1.

Note: This project, including this readme, was vibe coded with Windsurf / Sonnet 3.7 and GitHub Copilot Agent / GPT 4.1.

Demo TextTV Mode

Recording of the TextTV viewer navigating pages.

Demo URL Mode

Recording of the TextTV viewer navigating to a url.

Project Overview

This application offers two main modes of operation:

1. TextTV Mode: Fetches TextTV content from the texttv.nu API
2. URL Mode: Converts web page content to TextTV format using OpenAI's GPT-4.1

Both modes render the content with the classic TextTV styling:

• Yellow headlines
• Cyan secondary text
• White regular text
• Blue navigation bar at the bottom

Project Structure

The solution follows a clean architecture with separation of concerns:

TextTv.sln
├── TextTv.Cli (Main console application)
│   ├── Configuration/
│   │   ├── CommandLineOptions.cs (Command-line argument handling)
│   │   ├── PromptConstants.cs (OpenAI prompts for URL mode)
│   │   ├── Settings.cs (Application settings model)
│   │   └── SettingsProvider.cs (Settings file handling)
│   ├── Models/
│   │   └── TextTvPage.cs (Data model for API responses)
│   ├── Services/
│   │   ├── TextTvService.cs (TextTV API interaction)
│   │   ├── OpenAi/
│   │   │   └── OpenAiService.cs (OpenAI API integration)
│   │   └── Web/
│   │       └── WebContentService.cs (Web content fetching)
│   ├── Rendering/
│   │   └── TextTvRenderer.cs (Console output formatting)
│   └── Helpers/
│       ├── TextParsingHelper.cs (HTML parsing utilities)
│       └── Utils.cs (General utility methods)
└── TextTv.Cli.Tests (Unit tests)
    ├── HttpClientMock.cs (HTTP client mocking)
    ├── TextParsingHelperTests.cs
    ├── TextTvServiceTests.cs
    └── UtilsTests.cs

Key Design Decisions

1. Separation of Concerns:

   • The code is organized into logical areas (Models, Services, Rendering, Helpers, Configuration) to improve maintainability and readability.
   • Each class has a clear single responsibility.

2. Multi-Mode Operation:

   • TextTV Mode: Original functionality to fetch content from the texttv.nu API.
   • URL Mode: New functionality to convert web content to TextTV format using OpenAI.
   • Consistent rendering interface for both modes.

3. Clean API Interaction:

   • Abstracted API calls through dedicated service classes.
   • Well-defined model with TextTvPage record shared across modes.
   • Consistent error handling for both API types.

4. HTML Parsing and Rendering:

   • Using HtmlAgilityPack for robust HTML parsing.
   • The TextParsingHelper handles all HTML parsing logic.
   • Mode-aware TextTvRenderer for different UI presentations.

5. AI Integration:

   • Carefully designed prompts for optimal TextTV format conversion.
   • Few-shot learning examples to ensure high-quality output.
   • Content limitations to fit TextTV screen dimensions.

6. Error Handling:

   • Robust error handling for API requests, JSON parsing, and rendering.
   • Graceful fallbacks when parsing fails.

7. Testing:

   • Comprehensive unit tests for all components.
   • Mock HTTP client for testing API interactions without actual network calls.

Dependencies

• .NET 9.0
• HtmlAgilityPack (for HTML parsing)
• OpenAI (official .NET SDK for OpenAI integration)
• Microsoft.Extensions.Configuration.Json (for settings handling)
• Moq (for testing)
• MSTest framework

Building the Application

To build the application:

cd src/
dotnet build

Running the Application

Auto-Detection Mode (New!)

The application can now automatically detect if the input is a URL or a page number:

cd src/texttv-csharp/TextTv.Cli
# For page numbers
dotnet run -- 100

# For URLs
dotnet run -- https://example.com

This simplifies usage by automatically determining the appropriate mode based on your input.

TextTV Mode

You can still explicitly specify TextTV mode using the --pagenumber flag:

dotnet run -- --pagenumber 100

Replace 100 with any valid Text TV page number.

URL Mode

Similarly, you can explicitly use URL mode with the --url flag:

dotnet run -- --url https://example.com

Replace https://example.com with any valid URL you want to convert to TextTV format.

Setting Up OpenAI API Key

To use the URL mode, you need to configure your OpenAI API key:

1. Make sure you have an OpenAI API key. If you don't have one, you can create it in your OpenAI dashboard.

2. Create an appsettings.json file in the TextTv.Cli directory with the following content:

{
  "OpenAiApiKey": "your-api-key-here",
  "OpenAiModel": "gpt-4.1"
}

3. Replace your-api-key-here with your actual OpenAI API key.

Note: The appsettings.json file is excluded from git via .gitignore to prevent accidental exposure of your API key.

Publishing as a Single Executable

To publish the application as a self-contained, single-file executable for Linux:

dotnet publish src/TextTv.Cli/TextTv.Cli.csproj -c Release -r linux-x64 -p:PublishSingleFile=true -p:AssemblyName=texttv --self-contained -o publish

This creates a self-contained executable in the publish folder that can run on any compatible Linux system without requiring .NET to be installed.

To run the published application:

cd publish
./TextTv.Cli 100

Running Tests

To run the unit tests:

cd src/TextTv.Cli.Tests
dotnet test

API Usage

The application uses the texttv.nu API with the following format:

https://texttv.nu/api/get/{pageNumber}?includePlainTextContent=1

This returns a JSON array containing page content and navigation information.