Added cffi dep note to README, added cffi and Cython version to Docke…

…rfile

Dockerfile

@@ -33,15 +33,17 @@ ARG NUMPY_VERSION=1.14.3
 # Locked to Pandas 0.20.3 by https://github.com/rapidsai/cudf/issues/118
 ARG PANDAS_VERSION=0.20.3
 ARG PYARROW_VERSION=0.10.0
+ARG CYTHON_VERSION=0.29.1
 RUN conda install -n cudf -y -c numba -c conda-forge -c nvidia -c rapidsai -c defaults \
       numba=${NUMBA_VERSION} \
       numpy=${NUMPY_VERSION} \
       pandas=${PANDAS_VERSION} \
       pyarrow=${PYARROW_VERSION} \
+      cython=${CYTHON_VERSION} \
       nvstrings \
       cmake=3.12 \
       gtest=1.8.0 \
-      cython \
+      cffi \
       pytest
 
 # Clone cuDF repo
@@ -58,9 +60,7 @@ RUN source activate cudf && \
     cmake .. -DCMAKE_INSTALL_PREFIX=${CONDA_PREFIX} && \
     make -j install && \
     make python_cffi && \
-    make install_python && \
-    cd /cudf/cpp/build/python && \
-    python setup.py install
+    make install_python
 
 # cuDF build/install
 RUN source activate cudf && \

README.md

@@ -106,8 +106,7 @@ $ make test
 Build and install cffi bindings:
 ```bash
 $ make python_cffi                                  # build CFFI bindings for librmm.so, libcudf.so
-$ make install_python                               # build python bindings
-$ cd python && python setup.py install              # install python bindings to site-packages folder
+$ make install_python                               # build & install CFFI python bindings. Depends on cffi package from PyPi or Conda
 $ py.test -v                                        # optional, run python tests on low-level python bindings
 ```