Indexing PDF with Solr - solr

Question

Can anyone point me to a tutorial.
My main experience with Solr is indexing CSV files. But I cannot find any simple instructions/tutorial to tell me what I need to do to index pdfs.
I have seen this: http://wiki.apache.org/solr/ExtractingRequestHandler
But it makes very little sense to me. Do I need to install Tika?
Im lost - please help

Answer 1

With solr-4.9 (the latest version as of now), extracting data from rich documents like pdfs, spreadsheets(xls, xlxs family), presentations(ppt, ppts), documentation(doc, txt etc) has become fairly simple.
The sample code examples provided in the downloaded archive from
here contains a basic solr template project to get you started quickly.
The necessary configuration changes are as follows:
Change the solrConfig.xml to include following lines :
<lib dir="<path_to_extraction_libs>" regex=".*\.jar" />
<lib dir="<path_to_solr_cell_jar>" regex="solr-cell-\d.*\.jar" />
create a request handler as follows:
<requestHandler name="/update/extract"
startup="lazy"
class="solr.extraction.ExtractingRequestHandler" >
<lst name="defaults" />
</requestHandler>
2.Add the necessary jars from the solrExample to your project.
3.Define the schema as per your needs and fire a query like :
curl "http://localhost:8983/solr/collection1/update/extract?literal.id=1&literal.filename=testDocToExtractFrom.txt&literal.created_at=2014-07-22+09:50:12.234&commit=true" -F "myfile=#testDocToExtractFrom.txt"
go to the GUI portal and query to see the indexed contents.
Let me know if you face any problems.

Answer 2

You could use the dataImportHandler. The DataImortHandle will be defined at the solrconfig.xml, the configuration of the DataImportHandler should be realized in an different XML config file (data-config.xml)
For indexing pdf's you could
1.) crawl the directory to find all the pdf's using the FileListEntityProcessor
2.) reading the pdf's from an "content/index"-XML File, using the XPathEntityProcessor
If you have the list of related pdf's, use the TikaEntityProcessor
look at this http://solr.pl/en/2011/04/04/indexing-files-like-doc-pdf-solr-and-tika-integration/ (example with ppt) and this Solr : data import handler and solr cell

Answer 3

The hardest part of this is getting the metadata from the PDFs, using a tool like Aperture simplifies this. There must be tonnes of these tools
Aperture is a Java framework for extracting and querying full-text content and metadata from PDF files
Apeture grabbed the metadata from the PDFs and stored it in xml files.
I parsed the xml files using lxml and posted them to solr

Answer 4

Use the Solr, ExtractingRequestHandler. This uses Apache-Tika to parse the pdf file. I believe that it can pull out the metadata etc. You can also pass through your own metadata.
Extracting Request Handler

Answer 5

public class SolrCellRequestDemo {
public static void main (String[] args) throws IOException, SolrServerException {
SolrClient client = new
HttpSolrClient.Builder("http://localhost:8983/solr/my_collection").build();
ContentStreamUpdateRequest req = new
ContentStreamUpdateRequest("/update/extract");
req.addFile(new File("my-file.pdf"));
req.setParam(ExtractingParams.EXTRACT_ONLY, "true");
NamedList<Object> result = client.request(req);
System.out.println("Result: " +enter code here result);
}
This may help.

Answer 6

Apache Solr can now index all sort of binary files like PDF, Words, etc ... check out this doc:
https://lucene.apache.org/solr/guide/8_5/uploading-data-with-solr-cell-using-apache-tika.html