A starting point for research code, scientific software, and packages maintained by gflab. Select Use this template > Create a new repository, choose gflab as the owner, and select public or private visibility as appropriate. Project maintainers can add collaborators directly.
After creating a repository, replace this README with the project description and update CITATION.cff with the project title, actual authors, and repository URL. Documentation and other default project materials use professional academic English. The example below demonstrates the repository structure and can be replaced or removed.
Python 3.10 or later is required. No third-party dependencies are needed. Run the following command from the repository root:
python3 examples/summarize.py --input examples/toy.csvExpected output: {"count": 4, "mean": 2.5}. The input contains four synthetic observations and does not represent clinical data or a research finding. GitHub Actions runs the same example; a successful run establishes that the example executes, not that a scientific result has been validated or reproduced.
For your project, document installation, data access, a minimal execution command, expected outputs, and relevant hardware requirements and approximate runtime. Add dependency files appropriate to the implementation language and environment.
Provide the publication title, link or DOI, and the corresponding release or tag. State explicitly when a manuscript has not yet been published. Maintain ongoing software development in the same repository while retaining the version associated with each publication.
Identify the primary maintainer, a GitHub contact, and a colleague who can assume maintenance. Indicate whether the project is under development, actively maintained, or retained for reproducibility.
Collaborators may commit directly for routine work. Use pull requests when discussion or review would be useful; small projects do not require a mandatory review process. External users can raise questions through Issues.
Describe which results can be reproduced and which depend on restricted data, external models, or specific computational resources. Do not upload patient-level data to GitHub, including private repositories. Provide access instructions for public, synthetic, or appropriately authorized data instead.
Store large datasets and model weights in suitable external repositories. Record their versions, access links, and checksums where needed to identify the exact artifacts used.
Lab open-source code and software packages use the Apache License, Version 2.0 (Apache-2.0) by default. This template is distributed under that license; see LICENSE. Retain this default for new projects and update the copyright attribution as appropriate. Preserve applicable third-party licenses and notices, and respect any different requirements imposed by upstream code or dependencies. Specify data and model licenses separately.
Copyright 2026 FengGao Lab contributors. Licensed under the Apache License, Version 2.0.
See the lab's Research Software Publishing Guide. Use the distribution channel appropriate to the software ecosystem, such as PyPI, CRAN, or GitHub Packages/GHCR.