Invite open PRs from slundberg/shap to contribute

[dsgibbons]

Latest PRs to invite

The authors of the following open PRs should be invited to contribute to this fork. Ticked items are considered done for the purposes of this issue.

✉️ indicates an email has been sent. Items without this emoji did not have an obvious contact email.
✅ indicates a PR has been raised.
❓ indicates that we don't intend to raise a PR without further input from the PR's author.
❌ indicates the PR is out-of-date and will not be considered.

• Remove usage of np.int, np.float and np.bool shap/shap#1890 ✉️ ✅
• added full batch mode and interaction for permutation explainer shap/shap#1897 ❓
• Use full backward hook for compatibility with torch >= 1.8.0. shap/shap#1927 ✅
• Fixed typo and variable name in Introductory notebook shap/shap#1931 ✉️ ❌
• PartitionShap: channel-first partitioning and channel SHAP plots.  shap/shap#1944 ❓
• changes: shap/shap#2006 ✉️❓
• DOC: fix several doc issues and missing API documentation shap/shap#2020 ✉️
• replace numpy aliases with python defaults shap/shap#2048 ✅
• Added encoding call-out in the imdb module shap/shap#2056 ❓
• Being NaN-aware in calculation of means and their respective variances shap/shap#2123 ✉️ ✅
• https://github.com/slundberg/shap/pull/2155❓
• fixed Tree.__call__ to support lightgmb: plain DataFrame needs to be passed to predict() shap/shap#2166 ✉️ ✅
• [BUGFIX] ignore_inds should not be tuple shap/shap#2169
• Enabeling to create a beeswarm figure without sum of other features. shap/shap#2225
• https://github.com/slundberg/shap/pull/2280❓
• SHAP on converted models from onnx-tensorflow shap/shap#2290
• FIX expose Kernel explainer shap/shap#2452 ✉️
• Fix shrinking to -inf and blowing up to inf of mask_aug and eyAdj_aug of KernelExplainer shap/shap#2473
• add LayerNorm layer to deep pytorch implementation shap/shap#2549
• Refactor (isort, black) shap/shap#2594 ✉️ ✅
• remove unused imports shap/shap#2595 ✉️ ✅
• update(beeswarm): return figure object if show=False shap/shap#2602 ✉️
• Fix kernel explainer save / load method shap/shap#2604 ✉️
• MAINT refactor feature_perturbation in Tree explainers shap/shap#2624 ✉️
• Fix Issue #2528 shap/shap#2641 ✅
• Fixed IPython deprecation warning shap/shap#2649 ✅
• causalml trees support shap/shap#2654
• Fix Install dependencies stage in tests pipeline shap/shap#2655 ✅
• Fix: Waterfall plot .base_values error #2140 shap/shap#2667 ✉️
• Fix a small typo in docs shap/shap#2672 ✉️ ✅
• [BUGFIX] Fixed TypeError: unsupported operand type(s) for +: 'NoneType' and 'str' shap/shap#2679 ✉️
• Fix json serialization error in force_plot shap/shap#2683 ❓
• partial_dependence remove shap_values initlize shap/shap#2696
• Fix shap.summary_plot to work with matplotlib 3.6.0 shap/shap#2697 ✉️ ❓
• Fix PySpark GBT Issue shap/shap#2700
• [Bugfix] Fix TypeError at force_plot: can only concatenate str (not float) to str shap/shap#2710 ✉️
• fixes typos shap/shap#2726 ✅
• Update README.md shap/shap#2748 ❓
• Update _kernel.py shap/shap#2769 ✉️ ✅
• Update models.py shap/shap#2792 ✉️ ❌
• add brackets to the shape argument of reshape when calling as np.resh… shap/shap#2799 ✉️
• #2806 waterfall color customisation shap/shap#2807 ✉️
• Pytorch DeepExplainer register_full_backward_hook fix shap/shap#2812
• Replace warnings.DeprecationWarning with just DeprecationWarning shap/shap#2830 ✅
• A fix to IndexError in shap.plot.barh due to large cohorts # shap/shap#2833
• [BUGFIX] IndexError: tuple index out of range in shap.plots.bar #2832 shap/shap#2834
• Correct xlim in force_matplotlib in cases where the signs of force are all the same shap/shap#2839
• introduce max_val parameter in image plot shap/shap#2848 ✉️❓
• Replace sample dataset on readme shap/shap#2873 ✉️❓
• Update _monitoring.py to prevent failure of monitoring_plot when sample size is <50. shap/shap#2877
• Fix a spelling error in IMDB dataset documents shap/shap#2885 ✅
• Fixed Visual Studio 2022 support for Nvidia GPU support (NVCC Compilation Error Fix) shap/shap#2890 ✅
• [BUGFIX] yticklabels in plots.waterfall() shap/shap#2912 ✉️ ✅
• Fix GradientExplainer when tabular data is passed, instead of image shap/shap#2915
• Fix the word nerual to neural in the README shap/shap#2924 ✅

[11301858]

Hi, I'm the contributor for the last pull request on your list. May I ask what the purpose of this fork is?

[dsgibbons]

@11301858 please see the discussion here and feel free to ask questions there. Assuming slundberg/shap is no longer active, I'm looking for developers to help maintain and improve this fork of shap.

[11301858]

Hello, I will be glad to contribute! Thank you for letting me know of this opportunity. Mr. Lundberg still seems to be active, but I support the purpose of your work. Perhaps we can keep Mr. Lundberg in the loop just so that he knows what we're doing. Thanks, Shreyan

…

On Wed, Jan 4, 2023 at 2:32 PM dsgibbons ***@***.***> wrote: Please see the discussion here <#1> and feel free to ask questions there. Assuming slundberg/shap is no longer active, I'm looking for developers to help maintain and improve this fork of shap. — Reply to this email directly, view it on GitHub <#2 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/APY5NX6WBNGNPJBGPWLGDYDWQX265ANCNFSM6AAAAAASMFZINY> . You are receiving this because you commented.Message ID: ***@***.***>

[ekrell]

Hello! Stumbled on this today, and I would love to see progress on the SHAP library and appreciate your effort to get this going.
I am the author of PartitionShap: channel-first partitioning and channel SHAP plots.
Let me know how I can help.

-- Evan Krell (ekrell@islander.tamucc.edu)

[dsgibbons]

Hi @ekrell, great to hear from you. I'm looking to establish a community of maintainers so that we can share the load long-term. Do you know other developers that would be interested in helping out with this project?

[prabathbr]

Hi Thanks for the invite, I have raised #100 PR on this community edition for issue "shap#2890"

[dsgibbons]

Now we have write access at slundberg/shap, this issue can be closed. Any unticked PRs can be addressed over there.