azure-function

Business Central Code Interpreter - Azure Function

A secure Python execution environment for the Business Central Code Interpreter extension. This function enables data analysis on your Business Central data without exposing sensitive information outside your Azure subscription.

Setup

1. Deploy with Visual Studio Code

   • Install VS Code and the Azure Functions extension
   • Clone this repository and open in VS Code
   • Use Azure Functions extension to deploy directly to Azure:
     • Click the Azure icon in the sidebar
     • Select "Deploy to Function App..."
     • Choose "Create new Function App in Azure..."
     • Select Python runtime (3.10+)
     • Set HTTP trigger with "Function" authentication level
     • Choose a name, region, and plan (Consumption or Premium)

2. Business Central API Access

   • Skip this step if you already have BC API access configured
   • Register app in Azure AD
   • Add Business Central API permissions (API.ReadWrite.All)
   • Create a client secret

3. Configure Environment Variables

   • In the Azure Portal, navigate to your Function App
   • Go to "Settings" > "Configuration"
   • Add the following Application settings:
     • BC_TENANT_ID: Your Azure AD tenant ID
     • BC_CLIENT_ID: Client ID of the app registration
     • BC_CLIENT_SECRET: Secret of the app registration
   • Click "Save" when finished

Usage

Endpoint: POST /api/execute

Request:

{
  "code": "# Your Python code here"
}

Response Format:

{
  "data": "Your data as text",
  "chart_images": ["base64encoded_image1", "base64encoded_image2"]
}

Example

# Get data from Business Central
data = get_bc_data("v2.0/companies(123456)/salesInvoices", "sandbox")

# Create the output structure
output = {
    "data": {
        "total_invoices": len(data["value"]),
        "total_amount": sum(item["totalAmountIncludingTax"] for item in data["value"])
    },
    "chart_images": []
}

Available Tools

• Business Central Access: get_bc_data(relative_url, environment)
• Data Analysis: pandas, numpy, matplotlib, scikit-learn, statsmodels
• Visualization: Generate charts with matplotlib and return as base64-encoded images

Security

• Sandboxed execution environment
• Restricted access to system resources
• Dangerous Python modules and functions are blocked
• All data processing happens within your Azure subscription

Troubleshooting

• Check Application Insights logs for detailed error messages
• Verify environment variables are correctly set
• For complex data, convert to simple types or strings before returning

Support

For support, please contact dmitry@katson.com.

---

Part of the Code Interpreter extension for Microsoft Dynamics 365 Business Central