Some feedback on improving `TypeEvalPy`

[mir-am]

Hi @ashwinprasadme and @Samzcodez,

First of all, great job! I have managed to run and evaluate Type4Py on the micro-benchmark with almost one command. The automation is great! That said, I have some comments that might help improve the user experience with TypeEvalPy, especially with the installation process.

• Currently, one needs to run setup.sh, i.e., a shell script, to install the benchmark. This is good, though, but in general, people in the community are a bit hesitant to run a shell script for security reasons. Therefore, I suggest to create a proper Python package with setup.py file which installs the required dependencies and adds a command to the user's environment, e.g., typeevalpy. Then, the user can run the benchmark with one command, say, typeevalpy run. Or even typeevalpy run type4py, if one wants to run only one benchmark. This way, people only install a Python package, which should be also platform-agnostic, at least for Linux distributions.
• I could not find .results folder containing the benchmark output. It'd also be helpful to log the path to the results folder so that the user can find it easily.
• After running hityperdl, I also encountered this error below, although I could see the benchmark results at the end of the run.

  b"Python is running inside a Docker container\n2023-09-08 14:11:01,441 - runner - INFO - Command returned non-zero exit status: [Errno 2] No such file or directory: '/tmp/micro-benchmark/python_features/builtins/switch/._main_INFERREDTYPES.json' for file: /tmp/micro-benchmark/python_features/builtins/switch/main.py\n"

In general, I like the TypeEvalPy benchmark and it is pretty easy to use it.

[ashwinprasadme]

Thanks @mir-am. I have now changed the README to install and run TypeEvalPy without the script. I have also added CLI based selection of tools.

Itseems the .results will be hidden in the file manager. To avoid confusion, I have renamed the .results folder to just results

The hityper error is from the tool.

[mir-am]

Thanks for addressing the feedback. The installation process is better now.