This is the artifact released in the ICSE'24 paper "Less is More? An Empirical Study on Configuration Issues in Python PyPI Ecosystem".

All the data are stored here and docker images are built in this folder.

The docker images are built here to download the packages in Task1. It contains a file Dockerfile.

All packages are downloaded to this folder.

The docker images are built here to validate the import statements. It also contains a dockerfile Dockerfile and two python files test_python2.py and test_python3.py. test_python2.py and test_python3.py are used to perform validation in Docker for Python 2 and Python 3, respectively.

For more details, please access the docker_workdir directory.

If you want to use PyCon, please follow the steps below:

To start the Task1 to download the packages, you are required to input the path of pypi_packages.json and popular_packages.json to the python file installation_check.py.

python installation_check.py

If you finish downloading the packages and finish the dependency_check to collect the dependencies of the releases, you should comment out the line

install(install_file_path,workdir)

and uncomment the line

delete(workdir, popular_packages_file_path)

and run the code again. For more details, please access the installation check directory.

To collect the dependencies of each release, use the command:

usage: collect_dependency.py [-h] [-d DOWNLOADPATH] [-p POPULAR_PACKAGES] 
                [-s PYPI_PACKAGES]

optional arguments:
-h, --help            show this help message and exit
-d DOWNLOADPATH, --downloadpath DOWNLOADPATH
                        Path of downloaded packages
-p POPULAR_PACKAGES, --popular-packages POPULAR_PACKAGES
                        Path of popular_packages.json
-s PYPI_PACKAGES, --pypi-packages PYPI_PACKAGES
                        Path of pypi_packages.json

It will read pypi_packages.json as a catalog, analyze the packages and versions in the DOWNLOADPATH, 
and finally update popular_packages.json with the details of packages (install status, required dependencies,
installed dependencies, etc).

An example of running it:
python collect_dependency.py -d docker_workdir/pypi_packages -p popular_packages.json -s pypi_packages.json

To check the conflict of dependencies, use the command:

usage: check_conflict.py [-h] [-j POPULAR_PACKAGES] [-s PYPI_PACKAGES]
                [-o OUTPUT_PATH]

optional arguments:
-h, --help            show this help message and exit
-j POPULAR_PACKAGES, --popular-packages POPULAR_PACKAGES
                        Path of popular_packages.json
-s PYPI_PACKAGES, --pypi-packages PYPI_PACKAGES
                        Path of pypi_packages.json
-o OUTPUT_PATH, --output-path OUTPUT_PATH
                        Path of output (is a directory)

It will give a collection of all the versions that meet the requirements (only analyze packages with the status of
'installed' or 'Missing External Source Code'), using pypi_packages.json as a catalog.

An example of running it:
python check_conflict.py -j popular_packages.json -s pypi_packages.json -o task4_output_dir

For more details, please access the dependency check directory.

If you finish above tasks, then you can run the code.

First, you should run import_validation_requirements.py to get the requirements.txt file of each release. you are required to input the path of popular_packages.json to the python file import_validation_requirements.py.

python import_validation_requirements.py

Then, run code import_validation_STATUS0:

python import_validation_STATUS0.py

For each of the following steps, you are required to run the code import_caculate_status.py to refresh the status of each package.

python import_caculate_status.py

Then, run the codes import_validation_STATUS1, import_validation_STATUS2, import_validation_STATUS3 and import_validation_STATUS4 in turn:

python import_validation_STATUS1.py
python import_validation_STATUS2.py
python import_validation_STATUS3.py
python import_validation_STATUS4.py

Note the file import_validation_STATUS2.py needs to run in a python2 environment.

For the metadata of verfied third-party libraries on PyPI, please check the vlibs/ folder.

For the JSON file, install_status: True indicates that the library release passes the dependency check of PyCon. validated: True indicates that the library release passes all checks of PyCon.

If you use PyConf in your work, please cite us:

@misc{peng2024more,
      title={Less is More? An Empirical Study on Configuration Issues in Python PyPI Ecosystem}, 
      author={Yun Peng and Ruida Hu and Ruoke Wang and Cuiyun Gao and Shuqing Li and Michael R. Lyu},
      year={2024},
      eprint={2310.12598},
      archivePrefix={arXiv},
      primaryClass={cs.SE}
}