| Author: | Timo Schmidt <timo-schmidt@gmx.net> |
|---|---|
| Description: | Crawler based on mapreduce for Apache Hadoop |
| Build status: |  |
Cloudcrawler is java based crawler based on hadoop/mapreduces. The goal was to get familar with apache hadoop.
Alternatives are:
- Apache Nutch
- Heretrix Crawler
You can compile a single jar file with the following command:
mvn compile assembly:single
The jar file is located in:
target/org.cloudcrawler-jar-with-dependencies.jar
The crawler can be used the following way:
1. Create a source file with crawling start urls (one per line) eg:
Copy the file to hdfs
hadoop fs -copyFromLocal crawl.txt /cloudcrawler/crawl/start/
Now the crawling can be starte:
hadoop org.cloudcrawler-jar-with-dependencies.jar crawl /cloudcrawler/crawl/start/ /cloudcrawler/crawl/out1/
The job can be repeated several times and the crawler will discover more and more pages
hadoop org.cloudcrawler-jar-with-dependencies.jar crawl /cloudcrawler/crawl/out1/ /cloudcrawler/crawl/out2/
hadoop org.cloudcrawler-jar-with-dependencies.jar crawl /cloudcrawler/crawl/out2/ /cloudcrawler/crawl/out3/
When the crawling is done, there are some other jobs that can be used one of the is the linktrust job. This job should be executed 3-4 times and the first input is the last output of the crawling process
hadoop org.cloudcrawler-jar-with-dependencies.jar linktrust /cloudcrawler/crawl/out3/ /cloudcrawler/linktrust/out1/
The last process ist the indexing process. The indexing process can be used to write the documents to elasticsearch or solr
hadoop org.cloudcrawler-jar-with-dependencies.jar index /cloudcrawler/linktrust/out1/ /cloudcrawler/index/out1/