Recent advancements in Large Language Models (LLMs) and their utilization in code generation tasks have significantly reshaped the field of software development. Despite the remarkable efficacy of code completion solutions in mainstream programming languages, their performance lags when applied to less ubiquitous formats such as OpenAPI definitions. This study evaluates the OpenAPI completion performance of GitHub Copilot, a prevalent commercial code completion tool, and proposes a set of task-specific optimizations leveraging Meta's open-source model Code Llama. A semantics-aware OpenAPI completion benchmark proposed in this research is used to perform a series of experiments through which the impact of various prompt-engineering and fine-tuning techniques on the Code Llama model's performance is analyzed. The fine-tuned Code Llama model reaches a peak correctness improvement of 55.2% over GitHub Copilot despite utilizing 25 times fewer parameters than the commercial solution's underlying Codex model. Additionally, this research proposes an enhancement to a widely used code infilling training technique, addressing the issue of underperformance when the model is prompted with context sizes smaller than those used during training. The dataset, the benchmark, and the model fine-tuning code are made publicly available.

I am profiling APIs as part of my partnership with Streamdata.io, and my continued API Stack work. As part of my work, I am creating OpenAPI, Postman Collections, and APIs.json indexes for APIs in a variety of business sectors, and as I'm finishing up the profile for ParallelDots machine learning APIs, I am struck (again) by the importance of tags within OpenAPI definitions when it comes to defining what any API does, and something that will have significant effects on the growing machine learning, and artificial intelligence space. While profiling ParallelDots, I had to generate the OpenAPI definition from the Postman Collection they provide, which was void of any tags. I went through the handful of API paths, manually adding tags for each of the machine learning resources. Trying to capture what resources were available, allowing for the discovery, filtering, and execution of each individual machine learning model being exposed using a simple web API.