In today’s environment, where tech and non-tech fields are getting closer, turning complex code into easy-to-understand business English is crucial. CodeTeller is a tool designed to do just that. If you are unfamiliar with the terms RAG, Semantic Kernel, and Knowledge Graph, please review my previous articles so you can get up to speed on the building blocks of this offering. Now, let’s dive into the mechanics of CodeTeller.
First, let’s talk about what a “plan” is. A plan is a set of steps designed to turn code into useful business insights. You can create these plans by hand using formats like handlebar or yaml, or let an LLM do it by feeding it commands. Imagine telling it to “read all files from a directory, translate them into business English, organize that into a knowledge graph, and then make a document from that graph to keep in a database.” This initiates our journey with a clear blueprint.
How CodeTeller Works: A Step-by-Step Guide
Step 1: Reading Code
CodeTeller kicks off by identifying which code files to examine, using a Native Function to read and prep them for the next steps.
Step 2: Explaining Code
An AI plugin then translates the code’s purpose into plain English, making it understandable for anyone not familiar with coding. This flexibility to handle different programming languages is one of CodeTeller’s strong suits.
Step 3: Building a Knowledge Graph
The translated text is used to build a knowledge graph. This involves using another AI plugin to figure out the key components and how they relate, organizing this information into a structured, visual format.
Step 4: Storing Knowledge
CodeTeller saves these insights from each file in a relational database. This way, it keeps a detailed record of how different pieces of code are linked.
This cycle (Steps 1-4) repeats for each piece of code in a project.
Step 5: Relating Business Entities
CodeTeller looks at how different parts of business entities are linked through the code, building a network of connections. This helps show how various parts of the code affect each other.
Step 6: Generating Summaries
It then creates summaries of these connections, forming a story that explains how everything works together. This is done through another AI plugin that stitches together a narrative.
Step 7: Documentation for End Users
Finally, these stories are saved in a vector database, which makes them easy to access. This gives a full picture of what the code means for the business, translating tech speak into business insights.
Steps 5-7 are repeated for each key part of the business entities.
Why It’s a Big Deal
CodeTeller aims to make tech easier to understand for everyone. By turning code into business English, it encourages better communication between those who build the tech and those who use it. This step-by-step method ensures that complex coding is broken down into clear, actionable insights.
Looking Ahead
This approach highlights how AI can help bridge the gap between different fields. Whether you’re coding and want to share your work, or you’re on the business side and need to understand tech better, CodeTeller opens up new paths for collaboration. This is just the beginning of making tech talk accessible to all.
Learn more about Generative AI development with First Line Software
Your Technological Solution Expert
Nikolay Moskalev
Solution Architect at First Line Software.
With more than 17 years of experience in software development and cloud architecture, Nikolay Moskalev is a seasoned Solution Architect who excels in team leadership. He has a deep passion for AI and a proven track record of enhancing client relations through direct, on-site collaboration. Nikolay specializes in crafting solutions that address both immediate and medium-term challenges faced by organizations. Holding a Master’s degree in Computer Science and Mathematics, his extensive knowledge and technical proficiency make him an invaluable asset to any team seeking to drive significant results through innovative technology solutions.
Clinovera
