updating some docs

README.md

@@ -46,6 +46,22 @@ That's it.  NetworkX now leverages cuGraph for accelerated graph algorithms.
 
 -----
 
+## NOTICES
+
+__Deprecaations__:
+ - Accepting and returning NetworkX objects is being deprecated and will be dropped before the end of 2024. Users are encouraged to use NetworkX directly with the new nx-cugraph backend plug-in.  
+See:  https://developer.nvidia.com/blog/accelerating-networkx-on-nvidia-gpus-for-high-performance-graph-analytics/
+
+__Repo Refactoring__:</br>
+We are moving away from this being a mono-repo.  Several graph packages will be moved into their own repo over the next few releases. This will not impact the final packages being produced, just where the source code lives.
+  - New nx-cugraph repo for the accelerated NetworkX backend software: nx-cugraph
+  - New cugraph-gnn repo to house our GNN works: cugraph-pyg, cugraph-dgl, and wholegraph
+  - New cugraph-docs repo to centralize documentation
+  - New cugraph-pg repo to house the property graph work which is currently in experimental
+
+
+----
+
 ## Table of contents
 - Installation
   - [Getting cuGraph Packages](./docs/cugraph/source/installation/getting_cugraph.md)

docs/cugraph/source/index.rst

@@ -18,7 +18,8 @@ RAPIDS Graph documentation
      - :abbr:`cugraph-service (Graph-as-a-service provides both Client and Server packages)`
    * - :abbr:`pylibcugraph (light-weight Python wrapper with no guard rails)`
      - :abbr:`cugraph-dgl (Accelerated extensions for use with the DGL framework)`
-     -
+     - :abbr:`nx-cugraph` (Accelerated backend for NetworkX)
+     - 
    * - :abbr:`libcugraph (C++ API)`
      - :abbr:`cugraph-pyg (Accelerated extensions for use with the PyG framework)`
      -
@@ -39,6 +40,11 @@ even CuPy or SciPy sparse Matrices.
 Note: We are redoing all of our documents, please be patient as we update
 the docs and links
 
+~~~~~~~~~~~~~~~~~~~~~
+Deprecaations NOTICES
+~~~~~~~~~~~~~~~~~~~~~
+ - Accepting and returning NetworkX objects is being deprecated and will be dropped before the end of 2024. Users are encouraged to use NetworkX directly with the new nx-cugraph backend plug-in.  
+See:  https://developer.nvidia.com/blog/accelerating-networkx-on-nvidia-gpus-for-high-performance-graph-analytics/
 |
 
 .. toctree::

docs/cugraph/source/installation/source_build.md

@@ -43,7 +43,7 @@ files](https://github.com/rapidsai/cugraph/blob/main/conda/environments).
 conda env create --name cugraph_dev --file $CUGRAPH_HOME/conda/environments/all_cuda-118_arch-x86_64.yaml
 
 # for CUDA 12.x
-conda env create --name cugraph_dev --file $CUGRAPH_HOME/conda/environments/all_cuda-12.2_arch-x86_64.yaml
+conda env create --name cugraph_dev --file $CUGRAPH_HOME/conda/environments/all_cuda-122_arch-x86_64.yaml
 
 
 # activate the environment