请求提供Windows 7系统下PredictionIO的分步安装指南

Hey there! I totally get how those Linux-centric commands in the Apache docs can feel overwhelming when you're just starting out with PredictionIO on Windows. Let's break down the installation process into simple, Windows-friendly steps that you can follow easily.

PredictionIO relies on a few tools to run properly—let's get those set up first:

• JDK 8: PredictionIO 0.12.1 requires Java 8 (not newer versions). Download either Oracle JDK 8 or OpenJDK 8, run the installer, then set up the JAVA_HOME environment variable pointing to your JDK folder (e.g., C:\Program Files\Java\jdk1.8.0_381). Add %JAVA_HOME%\bin to your system Path variable. Verify it works by opening Command Prompt (CMD) and running:

  java -version
  javac -version
  

  Both should show version 1.8.x.
• Scala 2.11.x: PredictionIO is built on Scala 2.11, so grab a 2.11.x release (like 2.11.12). Install it, set the SCALA_HOME environment variable to your Scala folder, and add %SCALA_HOME%\bin to your Path. Verify with:

  scala -version
  

• Apache Spark 2.1.1: Download the pre-built Spark 2.1.1 package for Hadoop 2.7 (look for spark-2.1.1-bin-hadoop2.7.tgz). Unzip it to a folder like C:\spark-2.1.1, set SPARK_HOME to this path, and add %SPARK_HOME%\bin to your Path. Test it by running:

  spark-shell
  

  If the Spark shell launches without errors, you're good to go.
• PostgreSQL: PredictionIO needs a database. Install PostgreSQL for Windows, create a database named pio (you can use the pgAdmin tool or SQL commands), and note your database username and password (e.g., user pio, password yourpassword).

The Apache docs use Linux tar commands, but we can handle this with Windows tools:

• Download the apache-predictionio-0.12.1.tar.gz file from the Apache PredictionIO archives.
• Use a tool like 7-Zip or WinRAR to extract the file twice: first extract the .tar from the .gz, then extract the contents of the .tar to a folder like C:\PredictionIO-0.12.1.
• Set the PIO_HOME environment variable to this folder, and add %PIO_HOME%\bin to your system Path.

Now let's tweak the configuration to match your Windows setup:

• Go to C:\PredictionIO-0.12.1\conf, copy the pio-env.sh.template file, and rename the copy to pio-env.sh.
• Open pio-env.sh in a text editor (like Notepad++) and update these key settings (use forward slashes for paths):

  # Set your Spark path
  export SPARK_HOME=C:/spark-2.1.1
  # Set your PostgreSQL connection string
  export POSTGRES_JDBC_URL=jdbc:postgresql://localhost:5432/pio?user=pio&password=yourpassword
  # Make sure JAVA_HOME is correctly set (if not already in system env)
  export JAVA_HOME=C:/Program Files/Java/jdk1.8.0_381
  

Let's check if everything is configured correctly:

• Open CMD and run:

  pio status
  

  This will scan your dependencies and configuration. If all checks pass (you'll see "All good!" messages), proceed.
• Next, create your first PredictionIO app (this sets up the database tables):

  pio app new MyFirstPredictionApp
  

  Save the App ID that's returned—you'll need it later when working with events and models.

The Event Server handles data ingestion for PredictionIO. Launch it by running:

pio eventserver

Wait a minute or two for it to start, then test it by visiting http://localhost:7070 in your browser. You should see a simple status page confirming the server is running.

Quick Note on GPG Verification #

The Apache docs mention gpg --verify to check the file's integrity. As a beginner, this is optional—you can skip it for now. If you want to do it later, install Gpg4win on Windows, import the Apache PredictionIO signing key, and run the verify command in a GPG-enabled terminal.

内容的提问来源于stack exchange，提问作者Nishit Oza