I already opened a PR in kabuki as well (https://github.com/hddm-devs/kabuki/pull/36). Basically, as of pandas 1.0 using .ix. has been removed, as .iloc and .loc are less ambiguous. I went through the hddm and kabuki code replacing .ix with the (hopefully) equivalent index methods.
As far as I can test it the code seems to run and I am not introducing any new errors in the tests. But, as I am new to the hddm and kabuki, I couldn't thoroughly test the changes. A simple example, however, seems to work with pandas 1.0.
Otherwise, setting a requirement of pandas <1.0 would of course also work.
Hi,
I already opened a PR in kabuki as well (https://github.com/hddm-devs/kabuki/pull/36). Basically, as of pandas 1.0 using
.ix.
has been removed, as.iloc
and.loc
are less ambiguous. I went through the hddm and kabuki code replacing.ix
with the (hopefully) equivalent index methods. As far as I can test it the code seems to run and I am not introducing any new errors in the tests. But, as I am new to the hddm and kabuki, I couldn't thoroughly test the changes. A simple example, however, seems to work with pandas 1.0. Otherwise, setting a requirement of pandas <1.0 would of course also work.Best, Simon