112 lines
2.8 KiB
JavaScript
112 lines
2.8 KiB
JavaScript
|
//Jan Holp, DP 2021
|
||
|
|
|
||
|
|
|
||
|
|
//client2 = cassandra
|
||
|
|
//client1 = elasticsearch
|
||
|
|
//-----------------------------------------------------------------
|
||
|
|
|
||
|
|
//require the Elasticsearch librray
|
||
|
|
const elasticsearch = require('elasticsearch');
|
||
|
|
const client1 = new elasticsearch.Client({
|
||
|
|
hosts: [ 'localhost:9200']
|
||
|
|
});
|
||
|
|
client1.ping({
|
||
|
|
requestTimeout: 30000,
|
||
|
|
}, function(error) {
|
||
|
|
// at this point, eastic search is down, please check your Elasticsearch service
|
||
|
|
if (error) {
|
||
|
|
console.error('Elasticsearch cluster is down!');
|
||
|
|
} else {
|
||
|
|
console.log('ELasticSearch is ok');
|
||
|
|
}
|
||
|
|
});
|
||
|
|
|
||
|
|
//create new index skweb2
|
||
|
|
client1.indices.create({
|
||
|
|
index: 'skweb2'
|
||
|
|
}, function(error, response, status) {
|
||
|
|
if (error) {
|
||
|
|
console.log(error);
|
||
|
|
} else {
|
||
|
|
console.log("created a new index", response);
|
||
|
|
}
|
||
|
|
});
|
||
|
|
|
||
|
|
//indexing method
|
||
|
|
const bulkIndex = function bulkIndex(index, type, data) {
|
||
|
|
let bulkBody = [];
|
||
|
|
id = 1;
|
||
|
|
const errorCount = 0;
|
||
|
|
/// dsdgfdhdhgdjd
|
||
|
|
data.forEach(item => {
|
||
|
|
bulkBody.push({
|
||
|
|
index: {
|
||
|
|
_index: index,
|
||
|
|
_type: type,
|
||
|
|
_id : id++, // id zmenit na url
|
||
|
|
}
|
||
|
|
});
|
||
|
|
bulkBody.push(item);
|
||
|
|
});
|
||
|
|
// console.log(bulkBody);
|
||
|
|
client1.bulk({body: bulkBody})
|
||
|
|
.then(response => {
|
||
|
|
|
||
|
|
response.items.forEach(item => {
|
||
|
|
if (item.index && item.index.error) {
|
||
|
|
console.log(++errorCount, item.index.error);
|
||
|
|
}
|
||
|
|
});
|
||
|
|
console.log(
|
||
|
|
`Successfully indexed ${data.length - errorCount}
|
||
|
|
out of ${data.length} items`
|
||
|
|
);
|
||
|
|
})
|
||
|
|
.catch(console.err);
|
||
|
|
};
|
||
|
|
|
||
|
|
|
||
|
|
const cassandra = require('cassandra-driver');
|
||
|
|
const client2 = new cassandra.Client({ contactPoints: ['localhost:9042'], localDataCenter: 'datacenter1', keyspace: 'websucker' });
|
||
|
|
const query1 = 'SELECT domain_name FROM websucker.domain_quality WHERE good_count > 0 ALLOW FILTERING';
|
||
|
|
//const query2 = 'SELECT * from websucker.content WHERE domain_name = ' + domain_name[i] + 'ALLOW FILTERING'; // body_size > 0
|
||
|
|
|
||
|
|
//-------------------------------------------------------------------------
|
||
|
|
|
||
|
|
var domain_name = [];
|
||
|
|
var object_list = [];
|
||
|
|
|
||
|
|
|
||
|
|
client2.execute(query1)
|
||
|
|
.then(result => {
|
||
|
|
let pole = result.rows.map(r => {
|
||
|
|
domain_name.push(r.domain_name)
|
||
|
|
});
|
||
|
|
console.log("Vsetky domenove mena : " , domain_name);
|
||
|
|
domain_name.forEach(name => {
|
||
|
|
let query = 'SELECT * from websucker.content WHERE domain_name = ' + "'" + name + "'" + ' and body_size > 0 ALLOW FILTERING';
|
||
|
|
client2.execute(query).then( res => {
|
||
|
|
res.rows.map(rr => {
|
||
|
|
object_list.push({
|
||
|
|
domain_name: rr.domain_name,
|
||
|
|
title: rr.title,
|
||
|
|
body: rr.body,
|
||
|
|
target_link : rr.target_link,
|
||
|
|
id : rr.target_link
|
||
|
|
});
|
||
|
|
//console.log(object_list);
|
||
|
|
bulkIndex('skweb2', 'web_page', object_list);
|
||
|
|
//console.log(object_list);
|
||
|
|
});
|
||
|
|
|
||
|
|
}).catch(error => console.log(error));
|
||
|
|
})
|
||
|
|
|
||
|
|
}).catch(err => console.log(err));
|
||
|
|
|
||
|
|
setTimeout(() => console.log(object_list, 15000));
|
||
|
|
|
||
|
|
|
||
|
|
|
||
|
|
|