Build a news recommendation engine based on PHP and coreseek

Building a news recommendation engine based on PHP and coreseek

Introduction:
With the rapid development of the Internet, the way people obtain information on a daily basis is also changing. How to quickly and accurately help users filter out news content that matches their interests has become an important challenge. In this article, we will introduce how to use PHP and coreseek to build a news recommendation engine based on keyword matching.

1. Engine architecture

The architecture of the news recommendation engine is shown in the figure below:

User--> Recommendation engine--> coreseek -- > News database

Users submit news keywords through the recommendation engine. The recommendation engine will pass the keywords to coreseek. Coreseek queries the matching news through the index database and returns it to the recommendation engine. The recommendation engine sorts and filters the returned news list and returns the results to the user.

2. Installing and configuring coreseek

First, we need to install and configure coreseek. coreseek is a Chinese full-text indexing tool based on the open source search engine Sphinx, which can be used for fast text retrieval. In the Linux environment, we can install coreseek through the following command:

wget http://www.coreseek.cn/uploads/csft/4.1/coreseek-4.1-beta.tar.gz
tar -zxvf coreseek-4.1-beta.tar.gz
cd coreseek-4.1-beta
./configure --prefix=/usr/local/coreseek
make && make install
cd /usr/local/coreseek
cp -r /usr/local/coreseek/mmseg-3.2.14/etc/* ./etc/
vi etc/csft.conf

In the csft.conf configuration file, we need to set the connection information of the news database, such as host name, port number, etc. .

3. Database and data import

Next, we need to create a news database and import news data. Assuming that we use MySQL as the database management system, we can create the database and tables through the following commands:

CREATE DATABASE news;
USE news;
CREATE TABLE news (
    id INT PRIMARY KEY AUTO_INCREMENT,
    title VARCHAR(255),
    content TEXT
);

Then, import the news data into the database:

INSERT INTO news (title, content) VALUES ('新闻标题1', '新闻内容1');
INSERT INTO news (title, content) VALUES ('新闻标题2', '新闻内容2');
...

After importing all the news data into the database, we need Set coreseek's index configuration fileetc/sphinx.conf：

source news
{
    type            = mysql
    sql_host        = localhost
    sql_user        = your_mysql_user
    sql_pass        = your_mysql_password
    sql_db          = news
    sql_port        = 3306

    sql_query       = SELECT id, title, content FROM news
}

index news_index
{
    source          = news
    path            = /usr/local/coreseek/var/data/news_index
    docinfo         = extern
    mlock           = 0
}

4. PHP code example

The following is a simple PHP code example, using To submit user keywords and obtain news recommendation results:

<?php
$keyword = $_GET['keyword'];

$sphinx = new SphinxClient();
$sphinx->SetServer('localhost', 9312);
$sphinx->SetMatchMode(SPH_MATCH_ALL);
$sphinx->SetLimits(0, 10);

$result = $sphinx->Query($keyword, 'news_index');

if ($result === false) {
    echo "查询失败";
} else {
    $ids = array_keys($result['matches']);
    $news = [];
  
    $pdo = new PDO('mysql:host=localhost;dbname=news', 'your_mysql_user', 'your_mysql_password');
    $stmt = $pdo->prepare("SELECT title, content FROM news WHERE id IN (" . implode(',', $ids) . ")");
    $stmt->execute();

    while ($row = $stmt->fetch(PDO::FETCH_ASSOC)) {
        $news[] = $row;
    }
  
    echo json_encode($news);
}
?>

In this example, we use the SphinxClient class provided by the sphinxapi extension library to query with coreseek. First, we set the host name and port number of coreseek through the SetServer method, then use the SetMatchMode method to set the matching mode (here is all matching), and finally use the Query method Submit user keywords for query.

If the query is successful, we can obtain the matching news id list through $result['matches'], and then use the PDO class to interact with MySQL to query the corresponding news title based on the id and content.

5. Conclusion

Through the above steps, we successfully built a news recommendation engine based on PHP and coreseek. You can perform secondary development according to your own needs, such as adding functions such as user login and personalized recommendations. I hope this article helps you build a news recommendation engine!

The above is the detailed content of Build a news recommendation engine based on PHP and coreseek. For more information, please follow other related articles on the PHP Chinese website!