interpretml
Soften version requirements
The pyproject.toml file pins several packages at specific versions, e.g., an older scikit-learn. This makes it complex to try out this library in an environment with other tools. Are the exact versions necessary? If not, perhaps they could be softened?
https://github.com/jiachangliu/FasterRisk/blob/82d4673946b20bf874106a43581e2670518cabb4/pyproject.toml#L21
We'd like to integrate fasterrisk in Orange Data Mining, and we have the same issue. We'd like to have pandas and scikit-learn opened up. If you lack time for that, should we make a PR? Thanks!
For the reference, here is the PR that used your library: biolab/orange3-prototypes/pull/264 Here is the draft of the blog post that describes it: https://orange-web2-git-fork-zanmervic-scoringsheet-biolab.vercel.app/blog/explain-scoring-sheet/
