A Unified Intermediate Representation for Infrastructure as code

This project is used to transform IaC templates into an intermediate representation. The IR can be used as a in-context learning strategy for LLMs to understand the IaC template better and generate a more compliant one. The IR can also be used to create smell detectors which each smell only need to be implemented once and applied to all supporting IaC languages.

About the files

file_loaded_result.json: A given IaC template will be loaded to python data structures and the file stores how an IaC template will be represented in python.
parser_result.json: The file stores the intermediate representation of an IaC template. dependency_graph.json: The raw data of the dependency graph which generated with the parser_result.json. dependency_graph.png: The image form of the dependency graph (generated with graphviz)

Support infrastructure as code language

1. CloudFormation
2. Terraform

Future update

1. Support more IaC languages (Pulumi, Azure Resource Manager, Bicep)