Release 1.30 of libpython-clj is out. We have: * Support for python classes in require-python. They can be required like modules although you have to understand that in that case you are essentially calling the class methods explicitly (meaning overridden methods don't work). This is what you would expect in Python also. * Deeper integration of numpy into the datatype math system meaning you can have a numpy object and type (+ numpy-ary 5) and get back a new numpy array. This uses the numpy functions under the cover. * A demonstration of a cutting edge facial rec engine (https://github.com/cnuernber/facial-rec) that uses docker, conda, python, and clojure all playing in the same box. * Lots of bugfixes. https://github.com/cnuernber/libpython-clj

Total kudos @chris441