Updates README file to include nx-cugraph user documentation, adds nx-cugraph to main README

README.md

@@ -35,8 +35,13 @@
 </div>
 
 -----
+## News
 
-## Table of content
+[nx-cugraph](./python/nx-cugraph/README.md) is a NetworkX backend that's available to provide GPU acceleration to NetworkX.
+
+-----
+
+## Table of contents
 - Installation
   - [Getting cuGraph Packages](./docs/cugraph/source/installation/getting_cugraph.md)
   - [Building from Source](./docs/cugraph/source/installation/source_build.md)
@@ -52,6 +57,7 @@
     - [External Data Types](./readme_pages/data_types.md)
   - [pylibcugraph](./readme_pages/pylibcugraph.md)
   - [libcugraph (C/C++/CUDA)](./readme_pages/libcugraph.md)
+  - [nx-cugraph](./python/nx-cugraph/README.md)
   - [cugraph-service](./readme_pages/cugraph_service.md)
   - [cugraph-dgl](./readme_pages/cugraph_dgl.md)
   - [cugraph-ops](./readme_pages/cugraph_ops.md)

python/nx-cugraph/README.md

@@ -2,31 +2,78 @@
 
 ## Description
 [RAPIDS](https://rapids.ai) nx-cugraph is a [backend to NetworkX](https://networkx.org/documentation/stable/reference/classes/index.html#backends)
-with minimal dependencies (`networkx`, `cupy`, and `pylibcugraph`) to run graph algorithms on the GPU.
+to run supported algorithms with GPU acceleration.
 
-### Contribute
+## System Requirements
 
-Follow instructions for [contributing to cugraph](https://github.com/rapidsai/cugraph/blob/branch-23.12
-and [building from source](https://docs.rapids.ai/api/cugraph/stable/installation/source_build/), then build nx-cugraph in develop (i.e., editable) mode:
-```
-$ ./build.sh nx-cugraph --pydevelop
-```
+Using nx-cugraph with this notebook requires the following:
+
+ * NVIDIA GPU, Pascal architecture or later
+ * CUDA 11.2, 11.4, 11.5, 11.8, or 12.0
+ * Python versions 3.9, 3.10, or 3.11
+ * NetworkX >= version 3.2
+
+More details about system requirements can be found in the [RAPIDS System Requirements documentation](https://docs.rapids.ai/install#system-req)..
+
+## Installation
 
-### Run tests
+nx-cugraph can be installed using either conda or pip.
 
-Run nx-cugraph tests from `cugraph/python/nx-cugraph` directory:
+### conda
 ```
-$ pytest
+conda install -c rapidsai-nightly -c conda-forge -c nvidia nx-cugraph
 ```
-Run nx-cugraph benchmarks:
+### pip
 ```
-$ pytest --bench
+python -m pip install nx-cugraph-cu11 --extra-index-url https://pypi.nvidia.com
 ```
-Run networkx tests (requires networkx version 3.2):
+Notes:
+
+ * Nightly wheel builds will not be available until the 23.12 release, therefore the index URL for the stable release version is being used in the pip install command above.
+ * Additional information relevant to installing any RAPIDS package can be found [here](https://rapids.ai/#quick-start).
+
+## Enabling nx-cugraph
+
+### `NETWORKX_AUTOMATIC_BACKENDS` environment variable.
+The `NETWORKX_AUTOMATIC_BACKENDS` environment variable can be used to have NetworkX automatically dispatch to specified backends an API is called that the backend supports.
+Set `NETWORKX_AUTOMATIC_BACKENDS=cugraph` to use nx-cugraph to GPU accelerate supported APIs with no code changes.
+Example:
 ```
-$ ./run_nx_tests.sh
+bash> NETWORKX_AUTOMATIC_BACKENDS=cugraph python my_networkx_script.py
 ```
-Additional arguments may be passed to pytest such as:
+
+### `backend=` keyword argument
+To explicitly specify a particular backend for an API, use the `backend=`
+keyword argument. This argument takes precedence over the
+`NETWORKX_AUTOMATIC_BACKENDS` environment variable. This requires anyone
+running code that uses the `backend=` keyword argument to have the specified
+backend installed.
+
+Example:
+```
+nx.betweenness_centrality(cit_patents_graph, k=k, backend="cugraph")
+```
+
+### Type-based dispatching
+
+NetworkX also supports automatically dispatching to backends associated with
+specific graph types. Like the `backend=` keyword argument example above, this
+requires the user to write code for a specific backend, and therefore requires
+the backend to be installed, but has the advantage of ensuring a particular
+behavior without the potential for runtime conversions.
+
+To use type-based dispatching with nx-cugraph, the user must import the backend
+directly in their code to access the utilities provided to create a Graph
+instance specifically for the nx-cugraph backend.
+
+Example:
 ```
-$ ./run_nx_tests.sh -x --sw -k betweenness
+import networkx as nx
+import nx_cugraph as nxcg
+
+G = nx.Graph()
+...
+nxcg_G = nxcg.from_networkx(G)             # conversion happens once here
+nx.betweenness_centrality(nxcg_G, k=1000)  # nxcg Graph type causes cugraph backend
+                                           # to be used, no conversion necessary
 ```