Updated based on nvliyuan's feedback

Signed-off-by: Sameer Raheja <sraheja@.nvidia.com>

Author: Sameer Raheja

README.md

@@ -1,9 +1,9 @@
 # NVIDIA cuDF plugin for Apache Spark
 NOTE: For the latest stable [README.md](https://github.com/nvidia/cudf-spark/blob/main/README.md) ensure you are on the main branch.
 
-The NVIDIA cuDF plugin for Apache Spark provides a plugin library [Apache
-Spark](https://spark.apache.org) that leverages GPUs to accelerate processing via the
-[cuDF](https://rapids.ai) (CUDA DataFrame) libraries.
+The NVIDIA cuDF plugin for [Apache Spark](https://spark.apache.org) provides a plugin library that
+leverages GPUs to accelerate processing via the [cuDF](https://github.com/rapidsai/cudf) (CUDA
+DataFrame) libraries.
 
 Documentation on the current release can be found [here](https://nvidia.github.io/cudf-spark/).
 

docs/additional-functionality/cache-serializer.md

@@ -1,6 +1,6 @@
 ---
 layout: page
-title: NVIDIA cuDF pluign for Apache Spark Cache Serializer
+title: NVIDIA cuDF plugin for Apache Spark Cache Serializer
 parent: Additional Functionality
 nav_order: 2
 ---

docs/additional-functionality/iceberg-per-table-scan-options.md

@@ -15,7 +15,7 @@ wants to tune scan behavior across an existing Spark session — for example
 before a benchmark run, or to try different split sizes against existing tables
 without modifying them.
 
-The NVIDIA cuDF plugin for Apachde Spark ships a thin Iceberg session-catalog wrapper that lets
+The NVIDIA cuDF plugin for Apache Spark ships a thin Iceberg session-catalog wrapper that lets
 these options be set via session conf at three scopes — per-table, per-catalog,
 and global. The wrapper is opt-in and a pure pass-through for any table whose
 conf is not set at any scope, so enabling it does not change behavior for

docs/dev/README.md

@@ -5,7 +5,7 @@ nav_order: 11
 has_children: true
 permalink: /developer-overview/
 ---
-# cuDF for Apache Spark Developer Overview
+# NVIDIA cuDF plugin for Apache Spark Developer Overview
 This document provides a developer overview of the project and covers the
 following topics:
 * [Spark SQL and Query Plans](#spark-sql-and-query-plans)
@@ -79,7 +79,7 @@ turn will need to pull values from the input nodes, chaining all the way down
 the tree until eventually the iterator of the leaf nodes is pulled and causes
 the reading of rows from the raw input data.
 
-## How the cuDF for Spark Plugin Works
+## How the cuDF Plugin Works
 The plugin leverages two main features in Spark.  The first is a
 plugin interface in [Catalyst](https://databricks.com/glossary/catalyst-optimizer)
 that allows the optimizer to be extended.  The plugin is a Catalyst extension
@@ -90,7 +90,7 @@ which allows extensions to operate on Spark SQL data in a `ColumnarBatch` form.
 Processing columnar data is much more GPU friendly than row-by-row processing.
 
 For example, the same query plan shown above becomes the following plan after
-being processed by the cuDF for Spark plugin:
+being processed by the cuDF plugin:
 ```
 *(5) Sort [o_orderpriority#5 ASC NULLS FIRST], true, 0
 +- Exchange rangepartitioning(o_orderpriority#5 ASC NULLS FIRST, 200), true, [id=#611]
@@ -176,7 +176,7 @@ scala> import com.nvidia.spark.rapids.RapidsConf
 import com.nvidia.spark.rapids.RapidsConf
 
 scala> RapidsConf.help(true)
-# cuDF for Spark Plugin Configuration
+# cuDF Plugin Configuration
 The following is the list of options that `rapids-plugin-4-spark` supports.
 
 On startup use: `--conf [conf key]=[conf value]`. For example:

docs/dev/shims.md

@@ -7,12 +7,11 @@ parent: Developer Overview
 
 # Shim Development
 
-cuDF for Apache Spark supports multiple feature version lines of
-Apache Spark such as 3.3.x, 3.4.x, 3.5.x, 4.x and a number of vendor releases that contain
-a mix of patches from different upstream releases. These artifacts are generally
-incompatible between each other, at both source code level and even more often
-at the binary level. The role of the Shim layer is to hide these issues from the
-common code, maximize reuse, and minimize logic duplication.
+The NVIDIA cuDF plugin for Apache Spark supports multiple feature version lines of Apache Spark such
+as 3.3.x, 3.4.x, 3.5.x, 4.x and a number of vendor releases that contain a mix of patches from
+different upstream releases. These artifacts are generally incompatible between each other, at both
+source code level and even more often at the binary level. The role of the Shim layer is to hide
+these issues from the common code, maximize reuse, and minimize logic duplication.
 
 This is achieved by using a ServiceProvider pattern. All Shims implement the same API,
 the suitable Shim implementation is loaded after detecting the current Spark build version

docs/index.md

@@ -6,7 +6,7 @@ permalink: /
 description: This site serves as a collection of documentation about the NVIDIA cuDF plugin for Apache Spark
 ---
 # Overview
-**If you are a customer looking for information on how to adopt cuDF for Apache Spark
+**If you are a customer looking for information on how to adopt the cuDF plugin
 for your Spark workloads, please go to our User Guide for more information: [link](https://docs.nvidia.com/spark-rapids/user-guide/latest/index.html).**
 
 The cuDF plugin leverages GPUs to accelerate processing via the