AI supported refactoring release 0.1.0 to 0.2.1

In the refactoring effort the following LLMs have been involved:

ChatGPT-4, ChatGPT-5
Claude Sonnet 4.5
Gemini3 Pro Preview
Grok 4 Fast Free

Mostly via parallel OpenRouter usage and mix and match of proposed modifications. We have used https://gitingest.com/ to partly try to provide the whole repository as a background for prompts and try general prompts such as "Implement https://github.com/WolfgangFahl/pyLoDStorage/issues/151" simplify endpoints.yaml by referencing prefix sets in prefixes.yaml and cut and paste of the issue github markup. The results have only been partly useable but costly in terms of the number of prompt repetitions needed to get a useful result. The total LLM usage cost was less than $35 during the whole refactoring period - the fact that Grok 4 Fast Free was available has probably cut the cost by halfe. There were cases where prompts had to be refined more than a dozen times and frustration kicked in how difficult it was to avoid the larger models to not to try to produce too much code and go off the rails while following their trained statistical parrot trails. It was hard to contain damage by hallucinations that kept leaking in and could only be detected by the teste cases in the CI.

The following commits might contain LLM generated code:

Snapquery/LLMCommits

AI supported refactoring release 0.1.0 to 0.2.1

AI supported refactoring release 0.1.0 to 0.2.1[edit]

snapquery[edit]

pyLoDStorage[edit]

snapquery[edit]

pyLoDStorage[edit]