py-why / dowhy

DoWhy is a Python library for causal inference that supports explicit modeling and testing of causal assumptions. DoWhy is based on a unified language for causal inference, combining causal graphical models and potential outcomes frameworks.
https://www.pywhy.org/dowhy
MIT License
6.99k stars 923 forks source link

Some [source] links in the docs do not work #981

Open kbattocchi opened 1 year ago

kbattocchi commented 1 year ago

For example, in the datasets submodule, the [source] links next to at least lalonde_dataset, partially_linear_dataset, and psid_dataset all point to anchors that do not exist on the corresponding datasets source page (and the methods do not appear to be defined within that source at all).

Perhaps relatedly, it looks like some numpy methods are also being re-exported in the datasets module - choice and random, the former of which looks very strange in the documentation here since it says it's new as of version 1.7.0 which makes sense for numpy but not for dowhy; these entries don't have source links at all.

amit-sharma commented 1 year ago

Thanks for flagging this @kbattocchi These datasets do exist in datasets.py. I suspect some kind of caching during docs build that is leading to these incorrect results. @bloebp any idea? The numpy.choice mixup is rather strange.

bloebp commented 1 year ago

Hey, sorry I missed this before.

It is indeed strange and I don't see an obvious reason for these things. In 0.10, the source links are working, but not on main.

When building the documentation there is, however, a long list of issues and warnings, maybe it is time to go through this one by one. I can give it a shot, but can't promise when (it will definitely take some time to fix all of it).