Metadata-Version: 2.4
Name: swesmith
Version: 0.0.7
Summary: The official SWE-smith package - A toolkit for generating software engineering training data at scale.
Author-email: John Yang <byjohnyang@gmail.com>
License: MIT License
        
        Copyright (c) 2025 John Yang
        
        Permission is hereby granted, free of charge, to any person obtaining a copy
        of this software and associated documentation files (the "Software"), to deal
        in the Software without restriction, including without limitation the rights
        to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
        copies of the Software, and to permit persons to whom the Software is
        furnished to do so, subject to the following conditions:
        
        The above copyright notice and this permission notice shall be included in all
        copies or substantial portions of the Software.
        
        THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
        IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
        FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
        AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
        LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
        OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
        SOFTWARE.
        
Project-URL: Documentation, https://github.com/SWE-bench/SWE-smith
Project-URL: Bug Reports, https://github.com/SWE-bench/SWE-smith/issues
Project-URL: Source Code, https://github.com/SWE-bench/SWE-smith
Project-URL: Website, https://swesmith.com
Keywords: nlp,benchmark,code
Classifier: Programming Language :: Python :: 3.10
Classifier: Programming Language :: Python :: 3.11
Classifier: Programming Language :: Python :: 3 :: Only
Classifier: License :: OSI Approved :: MIT License
Classifier: Operating System :: OS Independent
Requires-Python: >=3.10
Description-Content-Type: text/markdown
License-File: LICENSE
Requires-Dist: astor
Requires-Dist: datasets
Requires-Dist: docker
Requires-Dist: ghapi
Requires-Dist: libcst
Requires-Dist: litellm
Requires-Dist: matplotlib
Requires-Dist: modal
Requires-Dist: openai
Requires-Dist: pre-commit
Requires-Dist: python-dotenv
Requires-Dist: rich
Requires-Dist: sglang
Requires-Dist: sparklines
Requires-Dist: swebench
Requires-Dist: tiktoken
Requires-Dist: tqdm
Requires-Dist: tree-sitter>=0.25
Requires-Dist: tree-sitter-c
Requires-Dist: tree-sitter-cpp
Requires-Dist: tree-sitter-c-sharp
Requires-Dist: tree-sitter-go
Requires-Dist: tree-sitter-javascript
Requires-Dist: tree-sitter-java
Requires-Dist: tree-sitter-ruby
Requires-Dist: tree-sitter-rust
Requires-Dist: unidiff
Requires-Dist: textual
Requires-Dist: tree-sitter-php>=0.23.11
Provides-Extra: docs
Requires-Dist: mkdocs; extra == "docs"
Requires-Dist: mkdocs-material; extra == "docs"
Requires-Dist: mkdocs-glightbox; extra == "docs"
Requires-Dist: mkdocs-include-markdown-plugin; extra == "docs"
Requires-Dist: mkdocstrings[python]>=0.18; extra == "docs"
Requires-Dist: mike; extra == "docs"
Provides-Extra: test
Requires-Dist: pytest; extra == "test"
Requires-Dist: pytest-cov; extra == "test"
Dynamic: license-file

<p align="center">
  <a href="https://swesmith.com/">
    <img src="docs/assets/banner.png" style="height: 10em" alt="Kawhi the SWE-smith" />
  </a>
</p>

<br>

<div align="center">
<a href="https://www.python.org/">
  <img alt="Build" src="https://img.shields.io/badge/Python-3.10+-1f425f.svg?color=purple">
</a>
<a href="https://copyright.princeton.edu/policy">
  <img alt="License" src="https://img.shields.io/badge/License-MIT-blue">
</a>
<a href="https://badge.fury.io/py/swesmith">
  <img src="https://badge.fury.io/py/swesmith.svg">
</a>
<a href="https://arxiv.org/abs/2504.21798">
  <img src="https://img.shields.io/badge/arXiv-2504.21798-b31b1b.svg">
</a>
</div>

<hr />

SWE-smith is a toolkit for training [SWE-agents](https://github.com/SWE-agent/SWE-agent). You can:
* Turn any Github repository into a [SWE-gym](https://github.com/SWE-Gym/SWE-Gym).
* Create *unlimited* tasks (e.g., file localization, program repair, [SWE-bench](https://github.com/SWE-bench/SWE-bench)) for that repo.
* Train an LM to become a better SWE ([SWE-agent-LM-32B](https://huggingface.co/SWE-bench/SWE-agent-LM-32B)).

## ⚒️ Build Environments
If you're interested in turning a GitHub repository into a SWE-gym, install the package from [source](https://swesmith.com/getting_started/installation/).

> [!TIP]
> SWE-smith requires Docker to create execution environments. SWE-smith was developed and tested on Ubuntu 22.04.4 LTS.
> We do *not* plan on supporting Windows or MacOS.

You can then build a dataset for the repository by...
1. [Creating an environment](https://swesmith.com/guides/env_construction/#create-an-execution-environment)
2. [Synthesizing task instances](https://swesmith.com/guides/create_instances/)
3. [Keep tasks that break 1+ unit tests](https://swesmith.com/guides/harnesses/)
4. [Generating issue text for your tasks](https://swesmith.com/guides/issue_gen/)

## 🏋️ Train SWE-agent's
Training SWE-agent's using the [SWE-smith dataset](https://huggingface.co/datasets/SWE-bench/SWE-smith) is super simple.
```python
from swesmith.profiles import registry
from datasets import load_dataset
ds = load_dataset("SWE-bench/SWE-smith", split="train") # Loads all 52k task instances
for task in ds:
    rp = registry.get_from_inst(task)  # Get the RepoProfile for the task
    container = rp.get_container(task) # Returns pointer to a Docker container with the task initialized

    """TODO: Train!"""
```

SWE-smith has been used to
* Fine-tune Qwen 2.5 Coder into SWE-agent-LM-32B (A +32% jump on SWE-bench Verified!) using [SWE-agent](https://github.com/SWE-agent/SWE-agent) [[Tutorial](https://swesmith.com/guides/train_swe_agent/)]
* Perform GRPO style reinforcement learning using [SkyRL](https://github.com/NovaSky-AI/SkyRL)

## 💿 Resources
* [52k Task Instances](https://huggingface.co/datasets/SWE-bench/SWE-smith)
* [SWE-agent-LM-32B](https://huggingface.co/SWE-bench/SWE-agent-LM-32B); **40.2%** pass@1 on [SWE-bench Verified](https://huggingface.co/datasets/SWE-bench/SWE-bench_Verified)!
* [26k SWE-agent Trajectories](https://huggingface.co/datasets/SWE-bench/SWE-smith-trajectories), including the 5k SWE-agent-LM-32B was trained on.
* [250+ Environments](https://github.com/SWE-bench/SWE-smith-envs), one Docker image per repo represented in SWE-smith.

And there's more coming!

## 💫 Contributions
We're actively working on several follow ups!
Check out the [Contributing Guide](CONTRIBUTING.md) for more.

Contact Person: [John Yang](https://john-b-yang.github.io/), [Kilian Lieret](https://lieret.net)
(Email: [johnby@stanford.edu](mailto:johnby@stanford.edu))

## 🪪 License
MIT. Check `LICENSE` for more information.

## ✍️ Citation

```bibtex
@misc{yang2025swesmith,
  title={SWE-smith: Scaling Data for Software Engineering Agents}, 
  author={John Yang and Kilian Lieret and Carlos E. Jimenez and Alexander Wettig and Kabir Khandpur and Yanzhe Zhang and Binyuan Hui and Ofir Press and Ludwig Schmidt and Diyi Yang},
  year={2025},
  eprint={2504.21798},
  archivePrefix={arXiv},
  primaryClass={cs.SE},
  url={https://arxiv.org/abs/2504.21798}, 
}
```

## 📕 Our Other Projects
<div align="center">
  <a href="https://github.com/SWE-bench/SWE-bench"><img src="docs/assets/swebench_logo_text_below.svg" alt="SWE-bench" height="120px"></a>
  &nbsp;&nbsp;
  <a href="https://github.com/SWE-agent/SWE-agent"><img src="docs/assets/sweagent_logo_text_below.svg" alt="SWE-agent" height="120px"></a>
  &nbsp;&nbsp;
  <a href="https://github.com/SWE-agent/Mini-SWE-Agent"><img src="docs/assets/mini_logo_text_below.svg" alt="Mini-SWE-Agent" height="120px"></a>
  &nbsp;&nbsp;
  <a href="https://github.com/SWE-agent/SWE-ReX"><img src="docs/assets/swerex_logo_text_below.svg" alt="SWE-ReX" height="120px"></a>
  &nbsp;&nbsp;
  <a href="https://github.com/SWE-bench/sb-cli"><img src="docs/assets/sbcli_logo_text_below.svg" alt="sb-cli" height="120px"></a>
</div>
