Moedim.ACA.Sessions 1.0.0

Moedim.ACA.Sessions

The second letter in the Hebrew alphabet is the ב bet/beit. Its meaning is "house". In the ancient pictographic Hebrew it was a symbol resembling a tent on a landscape.

A C# library for Azure Container Apps dynamic sessions - enabling secure, isolated code execution for AI agents and applications.

🚀 Overview

Moedim.ACA.Sessions provides a robust C# wrapper around Azure Container Apps' serverless code interpreter sessions, designed specifically for integration with Agentic AI frameworks where AI agents need to execute code safely and securely. This library enables AI agents to:

• Execute Python code in isolated, sandboxed environments
• Create and manipulate files (including PowerPoint presentations, Excel files, charts, etc.)
• Process data and generate visualizations
• Install and use popular Python packages (NumPy, pandas, matplotlib, etc.)
• Maintain session state across multiple code executions

Hire me

Please send email if you consider to hire me.

Give a Star! ⭐

If you like or are using this project to learn or start your solution, please give it a star. Thanks!

Install

    dotnet add package Moedim.ACA.Sessions

Projects

• Moedim.ACA.Sessions - actual library
• Moedim.ACA.Sessions.Example - usage

Configuration

appsettings.json

{
  "ACASessions": {
    "AzureTokenProvider": {
      "RefreshBeforeMinutes": 5,
      "Scopes": [
        "https://dynamicsessions.io/.default"
      ]
    },
    "SessionsHttpClient": {
      "Endpoint": "https://{region}.dynamicsessions.io/subscriptions/{subscriptionId}/resourceGroups/{resourceGroup}/sessionPools/{sessionPoolName}",
      "ApiVersion": "2024-02-02-preview"
    }
  }
}

🔐 Authentication

Azure RBAC Requirements Your application needs the following roles on the session pool:

• Azure ContainerApps Session Executor
• Contributor

Using Azure CLI (Development)

az account get-access-token --resource https://dynamicsessions.io

💻 Usage Examples

Basic Code Execution

// Register all ACA Sessions services
services.AddACASessions();

public class DataAnalysisAgent
{
    private readonly ICodeInterpreter _codeInterpreter;

    public DataAnalysisAgent(ICodeInterpreter codeInterpreter)
    {
        _codeInterpreter = codeInterpreter;
    }

    public async Task<string> AnalyzeDataAsync(string csvData, string sessionId)
    {
        // Upload CSV data
        await _codeInterpreter.UploadFileAsync(new FileUploadRequest
        {
            SessionId = sessionId,
            FileName = "data.csv",
            FileContent = Encoding.UTF8.GetBytes(csvData)
        }, CancellationToken.None);

        // Analyze data and create visualization
        var analysisCode = """
            import pandas as pd
            import matplotlib.pyplot as plt
            import json

            # Load and analyze data
            df = pd.read_csv('/mnt/data/data.csv')

            # Generate summary statistics
            summary = {
                'total_rows': len(df),
                'columns': list(df.columns),
                'numeric_summary': df.describe().to_dict()
            }

            # Create visualization
            plt.figure(figsize=(10, 6))
            df.hist(bins=20, figsize=(12, 8))
            plt.suptitle('Data Distribution')
            plt.tight_layout()
            plt.savefig('/mnt/data/analysis_chart.png', dpi=300, bbox_inches='tight')

            # Save analysis results
            with open('/mnt/data/analysis_results.json', 'w') as f:
                json.dump(summary, f, indent=2)

            print("Analysis complete!")
            print(f"Dataset contains {len(df)} rows and {len(df.columns)} columns")
            """;

        var result = await _codeInterpreter.ExecuteAsync(new CodeExecutionRequest
        {
            SessionId = sessionId,
            Code = analysisCode,
            SanitizeInput = true
        }, CancellationToken.None);

        return result.ToString();
    }
}

References

• Serverless code interpreter sessions in Azure Container Apps