AWS Flow Corrector

Converts hardcoded ARNs in Amazon Connect contact flows to Terraform template variables. This enables deploying multiple isolated environments (dev, staging, prod) or per-developer flow replicas without ARN conflicts.

Use case: Deploy individual dev environments where each developer has their own isolated copy of flows (e.g., dev_a_flow_4, dev_b_flow_4) that reference their own resources, all managed through Terraform with a single template.

Before:

{
  "LambdaFunctionARN": "arn:aws:lambda:us-east-1:123456789:function:state-lookup",
  "ContactFlowId": "arn:aws:connect:us-east-1:123456789:instance/abc/contact-flow/xyz"
}

After:

{
  "LambdaFunctionARN": "${function_state_lookup}",
  "ContactFlowId": "${contact_flow_sample_flow}"
}

Use in Terraform:

resource "aws_connect_contact_flow" "my_flow" {
  instance_id = aws_connect_instance.main.id
  name        = "My Flow"
  type        = "CONTACT_FLOW"
  
  content = templatefile("${path.module}/flow.out.json", {
    function_state_lookup            = aws_lambda_function.state_lookup.arn
    contact_flow_sample_flow         = aws_connect_contact_flow.sample.arn
  })
}

Usage

Single file:

python3 main.py -i example_flow.json

Batch process directory:

python3 main.py --input-dir example_flows --output-dir output_flows

Preview changes:

python3 main.py -i example_flow.json --dry-run -v

Options

Input/Output:

• -i, --input-file - Single input JSON file
• --input-dir - Process all .json files in directory
• -o, --output-file - Output file (default: {input}.out.json)
• --output-dir - Output directory for batch processing

Behavior:

• -d, --dry-run - Preview without writing files
• -m, --update-metadata - Update Metadata section (default: Actions only)
• -p, --print-keys - Show ARN mappings and exit
• -v, --verbose - Detailed output

Requirements

Python 3.8+ (no external dependencies)