extractors	model	stream	functions
name linguist name project-facts	gpt-4	true

Background

Prompts that use linguist and a custom project facts tool to try to classify a project so that an AI can retrieve the most appropriate knowledge base for working in the project.

Running the tool

docker run --rm -it \
           -v /var/run/docker.sock:/var/run/docker.sock \
           --mount type=volume,source=docker-prompts,target=/prompts \
           --mount type=bind,source=$HOME/.openai-api-key,target=/root/.openai-api-key \
           --mount type=bind,source=/Users/slim/docker/labs-make-runbook/prompts,target=/my-prompts \
           --workdir /my-prompts \
           vonwig/prompts:latest run \
                                 --host-dir $PWD \
                                 --user $USER \
                                 --platform "$(uname -o)" \
                                 --prompts-dir "github:docker/labs-make-runbook?ref=main&path=prompts/project_type"

prompt system

You are an expert at looking at the contents of git projects and understanding what technologies are being used.

prompt user

{{#linguist}}

This project contains {{language}} code.

{{/linguist}}

Here is a list of files that are currently versioned in this project:

{{#project-facts.files}}

• {{.}} {{/project-facts.files}}

Use this list of files and the languages that we've detected in the project to figure out what kind of projects this is. It is okay if it appears to be a combination of more than one project type, but try to be as specific as possible.

When you have made your choice, output using the following application/json format.

{"context": {"project": ["type"]}}

The project type should be selected from one of the following categories:

• GoLang
• NPM
• Python
• Clojure

Output only the above json and nothing else.