launch

🚀 RepoLaunch

Turning Any Codebase into Testable Sandbox Environment

RepoLaunch addresses the bottleneck of setting up execution environments by automating the process through an LLM-based agentic workflow.

Launch Environment

Before getting started, please set your OPENAI_API_KEY and TAVILY_API_KEY environment variable. We use tavily for LLM search engine support.

We provide an example input file test-dataset.jsonl and a run config test-config.json to help you quickly go through the launch process.

cd launch
pip install -e .

python -m launch.run --config-path test-config.json

Data Schema

Input

For the input data used to set up the environment, we require the following fields:

Field	Description
repo	Full name of the repository
base_commit	Commit to check out
instance_id	Unique identifier of the instance
language	Main language of the repo
created_at	(Optional) Creation time of the instance, used to support time-aware environment setup

Run Config

For the run configuration file, the following fields are supported:

Field	Type	Description
llm_provider_name	string	Name of the LLM provider (e.g., "OpenAI", "AOAI")
print_to_console	boolean	Whether to print logs to console
model_config	dict	Configuration for the LLM model (contains model_name and temperature)
workspace_root	string	Workspace folder for one run
dataset	string	Path to the dataset file
instance_id	string	Specific instance ID to run, null to run all instances in the dataset
first_N_repos	integer	Limit processing to first N repos (-1 for all repos)
max_workers	integer	Number of parallel workers for processing
overwrite	boolean	Whether to overwrite existing results (false will skip existing repos)

Output

The output will be saved in {playground_folder}/{instance_id}/result.json and follows the structure below:

Field	Description
instance_id	Unique identifier of the instance
base_image	Docker base image
setup_commands	List of shell commands used to set up the environment
test_commands	List of shell commands used to run the tests
duration	Time taken to run the process (in minutes)
completed	Boolean indicating whether the execution completed successfully
exception	Error message or null if no exception occurred