readme, tispark: update TiSpark and enable sparkR and pyspark (#27)

* update tispark to 1.0

* add TiSparkR

* upgrade tispark version to 1.0.1

README.md

@@ -184,3 +184,12 @@ scala> spark.sql("select count(*) from lineitem").show
 |   60175|
 +--------+
 ```
+
+You can also access Spark with Python or R using the following commands:
+
+```
+docker-compose exec tispark-master /opt/spark/bin/pyspark
+docker-compose exec tispark-master /opt/spark/bin/sparkR
+```
+
+More documents about TiSpark can be found [here](https://github.com/pingcap/tispark).

tispark/Dockerfile

@@ -2,25 +2,40 @@ FROM anapsix/alpine-java:8
 
 ENV SPARK_VERSION=2.1.1 \
     HADOOP_VERSION=2.7 \
-    TISPARK_VERSION=0.1.0-SNAPSHOT \
+    TISPARK_VERSION=1.0.1 \
+    TISPARK_R_VERSION=1.1 \
+    TISPARK_PYTHON_VERSION=1.0.1 \
     SPARK_HOME=/opt/spark \
     SPARK_NO_DAEMONIZE=true \
     SPARK_MASTER_PORT=7077 \
     SPARK_MASTER_HOST=0.0.0.0 \
     SPARK_MASTER_WEBUI_PORT=8080
 
+ADD R /TiSparkR
+
 # base image only contains busybox version nohup and ps
 # spark scripts needs nohup in coreutils and ps in procps
 # and we can use mysql-client to test tidb connection
-RUN apk --no-cache add coreutils procps mysql-client python py-pip R \
-    && pip install pytispark==1.0.1 pyspark==2.1.2
+RUN apk --no-cache add \
+        coreutils \
+        mysql-client \
+        procps \
+        python \
+        py-pip \
+        R \
+    && pip install --no-cache-dir pytispark==${TISPARK_PYTHON_VERSION} \
+    && R CMD build TiSparkR \
+    && R CMD INSTALL TiSparkR_${TISPARK_R_VERSION}.tar.gz \
+    && rm -rf /TiSparkR_${TISPARK_R_VERSION}.tar.gz /TiSparkR
 
 RUN wget -q https://archive.apache.org/dist/spark/spark-${SPARK_VERSION}/spark-${SPARK_VERSION}-bin-hadoop${HADOOP_VERSION}.tgz \
     && tar zxf spark-${SPARK_VERSION}-bin-hadoop${HADOOP_VERSION}.tgz -C /opt/ \
     && ln -s /opt/spark-${SPARK_VERSION}-bin-hadoop${HADOOP_VERSION} ${SPARK_HOME} \
-    && wget -q http://download.pingcap.org/tispark-${TISPARK_VERSION}-jar-with-dependencies.jar -P ${SPARK_HOME}/jars \
+    && wget -q https://github.com/pingcap/tispark/releases/download/${TISPARK_VERSION}/tispark-core-${TISPARK_VERSION}-jar-with-dependencies.jar -P ${SPARK_HOME}/jars \
     && wget -q http://download.pingcap.org/tispark-sample-data.tar.gz \
     && tar zxf tispark-sample-data.tar.gz -C ${SPARK_HOME}/data/ \
     && rm -rf spark-${SPARK_VERSION}-bin-hadoop${HADOOP_VERSION}.tgz tispark-sample-data.tar.gz
 
+ENV PYTHONPATH=${SPARK_HOME}/python/lib/py4j-0.10.4-src.zip:${SPARK_HOME}/python:$PYTHONPATH
+
 WORKDIR ${SPARK_HOME}

tispark/R/DESCRIPTION

@@ -0,0 +1,11 @@
+Package: TiSparkR
+Type: Package
+Title: TiSpark for R
+Version: 1.1
+Author: PingCAP
+Maintainer: Novemser <novemser@gmail.com>
+Description: A shabby thin layer to support TiSpark in R language.
+License: Apache 2.0
+Copyright: 2017 PingCAP, Inc.
+Encoding: UTF-8
+LazyData: true

tispark/R/NAMESPACE

@@ -0,0 +1 @@
+exportPattern("^[[:alpha:]]+")

tispark/R/R/tisparkR.R

@@ -0,0 +1,41 @@
+#
+# Copyright 2017 PingCAP, Inc.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#      http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# See the License for the specific language governing permissions and
+# limitations under the License.
+#
+#
+
+# Title     : TiSparkR
+# Objective : TiSpark entry for R
+# Created by: novemser
+# Created on: 17-11-1
+
+# Function:createTiContext
+# Create a new TiContext via the spark session passed in
+#
+# @return A new TiContext created on session
+# @param session A Spark Session for TiContext creation
+createTiContext <- function(session) {
+  sparkR.newJObject("org.apache.spark.sql.TiContext", session)
+}
+
+# Function:tidbMapDatabase
+# Mapping TiContext designated database to `dbName`.
+#
+# @param tiContext TiSpark context
+# @param dbName Database name to map
+# @param isPrefix Whether to use dbName As Prefix
+# @param loadStatistics Whether to use statistics information from TiDB
+tidbMapDatabase <- function(tiContext, dbName, isPrefix=FALSE, loadStatistics=TRUE) {
+  sparkR.callJMethod(tiContext, "tidbMapDatabase", dbName, isPrefix, loadStatistics)
+  paste("Mapping to database:", dbName)
+}

tispark/R/README.md

@@ -0,0 +1,44 @@
+## TiSparkR
+TiSparkR is a thin layer built to support the R language with TiSpark.
+
+### Usage
+1. Download the TiSparkR source code and build a binary package (run `R CMD build R` in TiSpark root directory). Install it to your local R library (e.g. via `R CMD INSTALL TiSparkR_1.0.0.tar.gz`)
+
+2. Build or download TiSpark dependency jar `tispark-core-1.0-RC1-jar-with-dependencies.jar` [here](https://github.com/pingcap/tispark).
+
+3. `cd` to your Spark home directory, and run:
+    ```
+    ./bin/sparkR --jars /where-ever-it-is/tispark-core-${version}-jar-with-dependencies.jar
+    ```
+    Note that you should replace the `TiSpark` jar path with your own.
+
+4. Use as below in your R console:
+    ```R
+    # import tisparkR library
+    > library(TiSparkR)
+    # create a TiContext instance
+    > ti <- createTiContext(spark)
+    # Map TiContext to database:tpch_test
+    > tidbMapDatabase(ti, "tpch_test")
+
+    # Run a sql query
+    > customers <- sql("select * from customer")
+    # Print schema
+    > printSchema(customers)
+    root
+     |-- c_custkey: long (nullable = true)
+     |-- c_name: string (nullable = true)
+     |-- c_address: string (nullable = true)
+     |-- c_nationkey: long (nullable = true)
+     |-- c_phone: string (nullable = true)
+     |-- c_acctbal: decimal(15,2) (nullable = true)
+     |-- c_mktsegment: string (nullable = true)
+     |-- c_comment: string (nullable = true)
+
+    # Run a count query
+    > count <- sql("select count(*) from customer")
+    # Print count result
+    > head(count)
+      count(1)
+    1      150
+    ```