Swebench

Latest version: v2.1.7

Safety actively analyzes 688634 Python packages for vulnerabilities to keep your Python projects secure.

Scan your dependencies

Page 2 of 3

1.1.5

* Add support for HumanEvalFix (Python, JS, Go, Java) ([source](https://huggingface.co/datasets/bigcode/humanevalpack))

1.1.0

* Add `env_vars_test` field to allow for environment variable assignment for testing scripts.
* Change `pip_packages` installation specification to be a list instead of a string.
* Define PyPI package versioning explicitly for dev, test repos.
* Fix versioning for `astroid` dependency in `pylint` installation script`.
* Fix minor error in `parse_log_pytest_options`.
* Improve clarity + succinctness of logging.
* Make logging of subprocess args to log file smaller.
* Remove installation specifications for `dbt-core`, `transformers`.
* Remove redundant declaration of constants.
* Remove unused versions from installation specifications for dev, test repos.
* Rewrite `swebench.metrics.get_model_report`.

1.0.5

* Fix log parsing for `pydicom`, `pylint`, and `requests` libraries. [5cb448](https://github.com/princeton-nlp/SWE-bench/commit/5cb448140a8cd05490650b0671d860765180f26c)

1.0.4

* Fixed `env_list` parsing. [5be59d](https://github.com/princeton-nlp/SWE-bench/commit/5be59d665233ffb63b9beb30b2740cc41098e51f)
* Updated `ExecWrapper`, `LogWrapper` logic for harness. [231a2b](https://github.com/princeton-nlp/SWE-bench/commit/231a2b205c5ca9ddcb126b73b22667d79e1b6108)

1.0.2

* Added `try/catch` around `lsof` based clean up for `run_evaluation.py`. [3fb217](https://github.com/princeton-nlp/SWE-bench/commit/3fb2179a5c69737465f916898e8708adffff9914)
* Fixed `get_eval_refs` function. [12a287](https://github.com/princeton-nlp/SWE-bench/commit/12a287a9591cb4a0d65483f0c8bfaa3375285bfc)
* Fixed `seaborn` log parser. [0372b6](https://github.com/princeton-nlp/SWE-bench/commit/0372b6a9ff62516067fb26f602163c231d818163)

1.0.1

First working version. We strongly recommend not using versions older than this one.
* Added logging for failed installations. [58d24d](https://github.com/princeton-nlp/SWE-bench/commit/58d24d1b65b95ed96d57805604aca7adca49861d)
* Added missing `datasets` dependency. [68e89e](https://github.com/princeton-nlp/SWE-bench/commit/68e89ef8d099ca5c23a8fd5681e3f990cf729fd6)
* Reorganized repository to be directly build-able as a PyPI package. [548bdb](https://github.com/princeton-nlp/SWE-bench/commit/548bdbffb2ac5f0a09c1d7eb95bbee1bce126233)

Page 2 of 3

© 2024 Safety CLI Cybersecurity Inc. All Rights Reserved.