Please also have a look at the Apache Nutch repository and all information about Apache Nutch given below.
Notable additions in Common Crawl's fork of Nutch (not yet pushed to upstream Nutch although this is planned):
- WARC and CDX writer integrated into Fetcher and able to detect the language of HTML pages using the CLD2 language detector
- Generator2: alternative implementation of Generator
- allowing to combine per-domain and per-host limits and
- optimized to create many (eg. 100) segments in a single job
How to install additional requirements to build this fork of Nutch:
- crawler-commons development snapshot package:
git clone [email protected]:crawler-commons/crawler-commons.git cd crawler-commons/ mvn install - install the latest public suffix list into
conf/to ensure that it is definitely used (see #17):wget https://publicsuffix.org/list/public_suffix_list.dat -O conf/effective_tld_names.dat - Java wrapper for CLD2 language detection
For runtime, if WARC language detection is enabled (
git clone [email protected]:commoncrawl/language-detection-cld2.git cd language-detection-cld2/ mvn installwarc.detect.language= true), also the CLD2 shared objects are required, e.g. on Ubuntusudo apt install libcld2-0 libcld2-dev
For the latest information about Nutch, please visit the Nutch website at:
and our wiki, at:
https://cwiki.apache.org/confluence/display/NUTCH/Home
To get started using Nutch read Tutorial:
https://cwiki.apache.org/confluence/display/NUTCH/NutchTutorial
To contribute a patch, follow these instructions (note that installing Hub is not strictly required, but is recommended).
- Download and install hub.github.com
- File JIRA issue for your fix at https://issues.apache.org/jira/projects/NUTCH/issues
- you will get issue id NUTCH-xxxx where xxxx is the issue ID.
git clone https://github.com/apache/nutch.gitcd nutchgit checkout -b NUTCH-xxxx- edit files (please try and include a test case if possible)
git status(make sure it shows what files you expected to edit)- Make sure that your code complies with the Nutch codeformatting template, which is basially two space indents
git add <files>git commit -m "fix for NUTCH-xxx contributed by <your username>"hub fork(if hub is not installed, you can fork the project using the "fork" button on the Nutch Github project page)git push -u <your git username> NUTCH-xxxxhub pull-request(if hub is not installed, please follow the instructions how to create a pull-request from a fork)
Generate Eclipse project files
ant eclipse
and follow the instructions in Importing existing projects.
You must configure the nutch-site.xml before running. Make sure, you've added http.agent.name and plugin.folders properties. The plugin.folders normally points to <project_root>/build/plugins.
Now create a Java Application Configuration, choose org.apache.nutch.crawl.Injector, add two paths as arguments. First one is the crawldb directory, second one is the URL directory where, the injector can read urls. Now run your configuration.
If we still see the No plugins found on paths of property plugin.folders="plugins", update the plugin.folders in the nutch-default.xml, this is a quick fix, but should not be used.
First install the IvyIDEA Plugin. then run ant eclipse. This will create the necessary
.classpath and .project files so that Intellij can import the project in the next step.
In Intellij IDEA, select File > New > Project from Existing Sources. Select the nutch home directory and click "Open".
On the "Import Project" screen select the "Import project from external model" radio button and select "Eclipse".
Click "Create". On the next screen the "Eclipse projects directory" should be already set to the nutch folder.
Leave the "Create module files near .classpath files" radio button selected.
Click "Next" on the next screens. On the project SDK screen select Java 11 and click "Create".
N.B. For anyone on a Mac with a homebrew-installed openjdk, you need to use the directory under libexec: <openjdk11_directory>/libexec/openjdk.jdk/Contents/Home.
Once the project is imported, you will see a popup saying "Ant build scripts found", "Frameworks detected - IvyIDEA Framework detected". Click "Import". If you don't get the pop-up, I'd suggest going through the steps again as this happens from time to time. There is another Ant popup that asks you to configure the project. Do NOT click "Configure".
To import the code-style, Go to Intellij IDEA > Preferences > Editor > Code Style > Java.
For the Scheme dropdown select "Project". Click the gear icon and select "Import Scheme" > "Eclipse XML file".
Select the eclipse-format.xml file and click "Open". On next screen check the "Current Scheme" checkbox and hit OK.
Running in Intellij
- Open Run/Debug Configurations
- Select "+" to create a new configuration and select "Application"
- For "Main Class" enter a class with a main function (e.g. org.apache.nutch.indexer.IndexingJob).
- For "Program Arguments" add the arguments needed for the class. You can get these by running the crawl executable for your job. Use full-qualified paths. (e.g. /Users/kamil/workspace/external/nutch/crawl/crawldb /Users/kamil/workspace/external/nutch/crawl/segments/20221222160141 -deleteGone)
- For "Working Directory" enter "/Users/kamil/workspace/external/nutch/runtime/local".
- Select "Modify options" > "Modify Classpath" and add the config directory belonging to the "Working Directory" from the previous step (e.g. /Users/kamil/workspace/external/nutch/runtime/local/conf). This will allow the resource loader to load that configuration.
- Select "Modify options" > "Add VM Options". Add the VM options needed. You can get these by running the crawl executable for your job (e.g. -Xmx4096m -Dhadoop.log.dir=/Users/kamil/workspace/external/nutch/runtime/local/logs -Dhadoop.log.file=hadoop.log -Dmapreduce.job.reduces=2 -Dmapreduce.reduce.speculative=false -Dmapreduce.map.speculative=false -Dmapreduce.map.output.compress=true)
Note: You will need to manually trigger a build through ANT to get latest updated changes when running. This is because the ant build system is separate from the Intellij one.