UchinaSystems.CsvToJson 0.1.3

UchinaSystems.CsvToJson (.NET)

High-performance CSV → JSON conversion powered by a Rust core via a small C ABI.

• Cross-platform native library loaded via runtimes/<rid>/native/
• Same options as the JS package: headers, delimiter, and header key case normalization

Install (from local build or NuGet)

• NuGet (once published):
  • Package Id: UchinaSystems.CsvToJson
• Or build locally and reference the project/package (.nupkg)

Quick Start

using System.Text;
using Csvjson;

var csv = "YouTube link,IMPORT\nhttps://yt.com,YES\n";
var bytes = Encoding.UTF8.GetBytes(csv);
var json = Parser.Parse(bytes, headers: true, delimiter: null, caseStyle: "pascal");
Console.WriteLine(json);

Result:

[{"YouTubeLink":"https://yt.com","Import":"YES"}]

API

namespace Csvjson
{
    public static class Parser
    {
        // Returns JSON string (array of objects)
        public static string Parse(
            byte[] bytes,
            bool? headers = null,   // default true when null
            byte? delimiter = null, // default 44 (',') when null
            string? caseStyle = null // one of: pascal|camel|snake|kebab|title|lower|upper
        );
    }
}

Case Normalization Examples

• "YouTube link" → "YouTubeLink" (pascal)
• "IMPORT" → "Import" (pascal)
• Mixed-case tokens like "YouTube" are preserved

Building Locally

This repo includes a Rust C ABI crate and a script to place the native library into the correct runtimes/ folder for packing.

# Build native lib for your platform and copy to runtimes/<rid>/native/
node scripts/build-dotnet.mjs

# Pack the NuGet package
dotnet pack ./dotnet/Csvjson/Csvjson.csproj -c Release

RIDs used:

• macOS: osx-x64, osx-arm64 → libcsvjson_cabi.dylib
• Linux: linux-x64, linux-arm64 → libcsvjson_cabi.so
• Windows: win-x64, win-arm64 → csvjson_cabi.dll

To ship multiple RIDs, repeat the native build for each platform (or cross-compile) and ensure the outputs exist under dotnet/Csvjson/runtimes/<rid>/native/ before packing.

Notes

• The parser is compute-only and pure; it doesn’t perform IO.
• For very large CSVs, prefer Node native add-on in server-side environments; the .NET wrapper is ideal for C# backends that want a simple drop-in conversion step.