A lightweight Python library to construct plain-text code prompts from your code bases, designed for use with Large Language Models (LLMs) in code analysis and assistance tasks.
code-prompt-builder simplifies the process of feeding code to your LLM. It reads files from your project, optionally filtering by file types and directories, and structures them into a single string suitable for LLM input.
This makes it easy to ask LLMs to analyze your code, identify potential issues, or even generate code suggestions.
pip install code-prompt-builderHere's a simple example of using code-prompt-builder:
from code_prompt_builder.utils import build_code_prompt
# Specify directories containing your code
repo_dir_paths = ["path/to/your/code/directory"]
# Create the code prompt
code_prompt = build_code_prompt(repo_dir_paths=repo_dir_paths)
# Now you can use 'code_prompt' as input to your LLM
print(code_prompt)repo_tree_paths: A list of root directory paths to walk through and collect files.repo_dir_paths: A list of directories to read files from directly (no recursive walk).repo_file_paths: A list of specific files to include.include_file_regex: (Optional) A regular expression to include only matching files.exclude_file_regex: (Optional) A regular expression to exclude matching files.