Development Environment Specification

[guyfreund]

Hi there, enjoyed reading your documentation, very elaborative!

I'm running Mac OSX 12.6 (M1) trying to set up a pipenv environment with the provided requirements file. I'm struggling to install all required versioned dependencies due to discrepancies. I found that problematic packages are arch, scikit-learn & statsmodels.

I was wondering if there might be an updated version of the dependencies versions. In addition, can you please elaborate on your python environment setup, e.g. what is the python versions manager (e.g. pyenv), and what tool do you use to handle python dependencies (e.g. pipenv).

Thanks in advance, Guy

[guyfreund]

I found out that the following requirement.txt file passes all tests.

mac os users with apple silicon needs to follow https://github.com/cvxopt/cvxopt/issues/214 and https://github.com/cvxopt/cvxopt/issues/214#issuecomment-1260989925 for the last step.

pandas==1.5.0 numpy==1.23.3 scipy==1.9.1 Pillow==9.2.0 MarkupSafe==2.1.1 matplotlib==3.6.0 joblib==1.2.0 scikit-learn==1.1.2 cvxopt==1.3.0 openpyxl==3.0.10 seaborn==0.12.0 weasyprint==56.1 emails==0.6 dic==1.5.2b1 Jinja2==3.1.2 arch==5.3.1 requests==2.28.1 statsmodels==0.13.2 xarray==2022.9.0 Quandl==3.7.0

@myrmarachne can we change it officially?

[unmais]

@guyfreund , your comment has helped me set my own environment - Python 3.10 under Ubuntu. This is the requirements.txt that made it for me (made the unit tests pass, no more than that):

pandas==1.5.2 numpy==1.23.5 scipy==1.9.3 Pillow==9.3.0 MarkupSafe==2.1.1 matplotlib==3.6.2 joblib==1.2.0 scikit-learn==1.1.3 cvxopt==1.3.0 openpyxl==3.0.10 seaborn==0.12.1 weasyprint==57.1 emails==0.6 dic==1.5.2b1 Jinja2==3.1.2 arch==5.3.1 requests==2.28.1 statsmodels==0.13.5 xarray==2022.11.0 Quandl==3.7.0

[myrmarachne]

Hi, we are adding support for python 3.10 and 3.11, along with the support for new dependencies in the next QF-Lib release.