Currently we just dump the doc contents (i.e. docstrings) into the output without any modification. This means that it doesn't really render nicely so I want to try and parse them in some capacity to see if we can render them better. This will probably be either markdown or rst and numpy or goolge style. We'll see what's appropriate. I do think that if the parsing fails we should just default back to dumping the plain contents, perhaps with a warning.
Currently we just dump the doc contents (i.e. docstrings) into the output without any modification. This means that it doesn't really render nicely so I want to try and parse them in some capacity to see if we can render them better. This will probably be either markdown or rst and numpy or goolge style. We'll see what's appropriate. I do think that if the parsing fails we should just default back to dumping the plain contents, perhaps with a warning.