Difference between revisions of "SMILA/5 Minutes Tutorial"

Revision as of 03:47, 9 April 2013

On this page we describe the necessary steps to install and run SMILA in order to create a search index on the SMILA Eclipsepedia pages and search them.

If you have any troubles or the results differ from what is described here, check the FAQ.

Supported Platforms

The following platforms are supported:

• Linux 32 Bit
• Linux 64 Bit
• Mac OS X 64 Bit (Cocoa)
• Windows 32 Bit
• Windows 64 Bit

Download and start SMILA

Download the SMILA package matching your operation system and unpack it to an arbitrary folder. This will result in the following folder structure:

/<SMILA>
  /configuration
  /features
  /jmxclient
  /plugins
  /workspace
  .eclipseproduct
  ...
  SMILA
  SMILA.ini

Preconditions

To be able to start SMILA, check the following preconditions first:

JRE

You will have to provide a JRE executable to be able to run SMILA. The JVM version should be Java 7. You may either:

• add the path of your local JRE executable to the PATH environment variable
  or
  
• add the argument -vm <path/to/jre/executable> right at the top of the file SMILA.ini.
  Make sure that -vm is indeed the first argument in the file, that there is a line break after it and that there are no leading or trailing blanks. It should look similar to the following:

-vm
d:/java/jre7/bin/java
...

Linux

When using the Linux distributable of SMILA, make sure that the files SMILA and jmxclient/run.sh have executable permissions. If not, set the permission by running the following commands in a console:

chmod +x ./SMILA
chmod +x ./jmxclient/run.sh

MacOS

When using MAC switch to SMILA.app/Contents/MacOS/ and set the permission by running the following command in a console:

chmod a+x ./SMILA

Start SMILA

To start the SMILA engine, simply double-click the SMILA executable. Alternatively, open a command line, navigate to the directory where you extracted the files to, and execute the SMILA executable. Wait until the engine has been fully started.

You can tell if SMILA has fully started if the following line is printed on the OSGI console: HTTP server started successfully on port 8080 and you can access SMILA's REST API at http://localhost:8080/smila/.

When using MAC, navigate to SMILA.app/Contents/MacOS/ in terminal, then start with ./SMILA

Before continuing, check the log file for possible errors.

Stop SMILA

To stop the SMILA engine, type close into the OSGI console and press Enter:

osgi> close

For further OSGI console commands, enter help:

osgi> help

Install a REST client

We're going to use SMILA's REST API to start and stop jobs, so you need a REST client. In REST Tools you find a selection of recommended browser plugins if you haven't got a suitable REST client yet.

Start Indexing Job and Crawl Import

Now we're going to crawl the SMILA Eclipsepedia pages and index them using the embedded Solr integration.

Start indexing job run

We are going to start the predefined indexing job "indexUpdate" based on the predefined asynchronous "indexUpdate" workflow. This indexing job will process the imported data.

Use your favorite REST Client to start a job run for the job "indexUpdate":

#Request
POST http://localhost:8080/smila/jobmanager/jobs/indexUpdate/

Your REST client will show a result like this:

#Response
{
  "jobId" : "20110901-121343613053",
  "url" : "http://localhost:8080/smila/jobmanager/jobs/indexUpdate/20110901-121343613053/"
}

You will need the "jobId" later on to finish the job run. The job run Id can also be found via the monitoring API for the job:

#Request
GET http://localhost:8080/smila/jobmanager/jobs/indexUpdate/

In the SMILA.log file you will see a message like that:

INFO ... internal.JobRunEngineImpl   - started job run '20110901-121343613053' for job 'indexUpdate'

Further information: The "indexUpdate" workflow uses the PipelineProcessorWorker that executes the synchronous "AddPipeline" BPEL workflow. So, the synchronous "AddPipeline" BPEL workflow is embedded in the asynchronous "indexUpdate" workflow. For more details about the "indexUpdate" workflow and "indexUpdate" job definitions see SMILA/configuration/org.eclipse.smila.jobmanager/workflows.json and jobs.json). For more information about job management in general please check the JobManager documentation.

Start the crawler

Now that the indexing job is running we need to push some data to it. There is a predefined job for indexing the SMILA Eclipsepedia pages which we are going to start right now. For more information about crawl jobs please see Importing Concept. For more information on jobs and tasks in general visit the JobManager manual.

To start the job run, POST the following JSON fragment with your REST client to SMILA:

#Request
POST http://localhost:8080/smila/jobmanager/jobs/crawlSmilaWiki/

This starts the job crawlSmilaWiki, which crawls the SMILA Eclipsepedia starting with http://wiki.eclipse.org/SMILA and (by applying the configured filters) following only links that have the same prefix. All pages crawled matching this prefix will be pushed to the import job.

If you like, you can monitor both job runs with your REST client at the following URIs:

• Crawl job: http://localhost:8080/smila/jobmanager/jobs/crawlSmilaWiki
• Import job: http://localhost:8080/smila/jobmanager/jobs/indexUpdate

Or both in one overview at

• http://localhost:8080/smila/jobmanager/jobs/

The crawling of the wikipedia page should take some time. If all pages are processed, the status of the crawlSmilaWiki's job run will change to SUCCEEDED. You can have a look at SMILA's search page to find out if some of the pages have already made their way into the Solr index.

Further information: You can find details about the relevant Import concepts here.

Search the index

Since SMILA uses Solr's autocommit feature (which is configured in solrconfig.xml to a period of 30 seconds or 1000 documents, whichever comes first) it might take some time until you retrieve results.

To search the index which was created by the crawlers, point your browser to http://localhost:8080/SMILA/search. There are currently two stylesheets from which you can select by clicking the respective links in the upper left corner of the header bar: The Default stylesheet shows a reduced search form with text fields like Query, Result Size, and Index, adequate to query the full-text content of the indexed documents. The Advanced stylesheet in turn provides a more detailed search form with text fields for meta-data search like for example Path, MimeType, Filename, and other document attributes.

To use the Default Stylesheet:

1. Point your browser to http://localhost:8080/SMILA/search.
2. Enter a word that you expect to be contained in your dummy files into the Query text field.
3. Click OK to send your query to SMILA.

To use the Advanced Stylesheet:

1. Point your browser to http://localhost:8080/SMILA/search.
2. Click Advanced to switch to the detailed search form.
3. For example, to find a file by its name, enter the file name into the Filename text field, then click OK to submit your search.

Stop indexing job run

Although there's no need for it, we can finish our previously started indexing job run via REST client now: (replace <job-id> with the job-id you got before when you started the job run).

#Request
POST http://localhost:8080/smila/jobmanager/jobs/indexUpdate/<job-id>/finish

You can monitor the job run via your browser to see that it has finished successfully:

#Request
GET http://localhost:8080/smila/jobmanager/jobs/indexUpdate/<job-id>

In the SMILA.log file you will see messages like this:

 INFO ... internal.JobRunEngineImpl   - finish called for job 'indexUpdate', run '20110901-141457584011'
 ...
 INFO ... internal.JobRunEngineImpl   - Completing job run '20110901-141457584011' for job 'indexUpdate' with final state SUCCEEDED

Congratulations, you've just crawled the SMILA Eclipsepedia, indexed the pages and searched through them. For more, just visit SMILA Manual.

Further steps

Crawl the filesystem

SMILA has also a predefined job to crawl the file system ("crawlFilesystem"), but you will have to either adapt the predefined job to point it to a valid folder in your filesystem or create your own job.

We will settle for the second option, because it does not need that you stop and restart SMILA.

Create your Job

POST the following job description to SMILA's Job API at http://localhost:8080/smila/jobmanager/jobs. Adapt the rootFolder parameter to point to an existing folder on your machine where you have placed some files (e.g. plain text, office docs or HTML files). If your path includes backslashes, escape them with an additional backslash, e.g. c:\\data\\files.

#Request
POST http://localhost:8080/smila/jobmanager/jobs/
{
  "name":"crawlFilesAtData",
  "workflow":"fileCrawling",
  "parameters":{
    "tempStore":"temp",
    "dataSource":"file",
    "rootFolder":"/data",
    "jobToPushTo":"indexUpdate",
    "mapping":{
      "fileContent":"Content",
      "filePath":"Path",       
      "fileName":"Filename",       
      "fileExtension":"Extension",
      "fileLastModified":"LastModifiedDate"
    }
  }
}

Hint: Not all file formats are supported by SMILA out-of-the-box. Have a look here for details.

Start your jobs

• Start the indexUpdate (see Start indexing job run), if you have already stopped it. If it is still running, that's fine:

#Request
POST http://localhost:8080/smila/jobmanager/jobs/indexUpdate/

• Start your crawlFilesAtData job similar to Start the crawler but now use the job name crawlFilesAtData instead of crawlSmilaWiki. This new job behaves just like the web crawling job, but its run time might be shorter, depending on how much data actually is at your rootFolder.

#Request
POST http://localhost:8080/smila/jobmanager/jobs/crawlFilesAtData/

Search for your new data

1. After the job run's finished, wait a bit, then check whether the data has been indexed (see Search the index for help).
2. It is also a good idea to check the log file for errors.

5 more minutes to change the workflow

The 5 more minutes to change the workflow show how you can configure the system so that data from different data sources will go through different workflows and pipelines and will be indexed into different indices.

(see Start indexing job run), if you have already stopped it. If it is still running, that's fine: