You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@jackrabbit.apache.org by Belinda Randolph <be...@jpl.nasa.gov> on 2007/05/02 22:42:23 UTC
JackRabbit Search Engine Questions
I am in the process of evaluating 10 repository solutions for my project.
I have several questions to ask in order to make my decisions.
1. Can I replace the JackRabbit search engine with my own?
2. Does your search engine look through actual document contents -
as a background process or at the time of the actual user search?
3. What FORMATs of actual documents does your search engine look
at? (Ascii, Microsoft, PDF, etc.)
4. When searching the contents of a PDF file, does the background
process, using OCR, create an additional file in another format? What format?
5. Does your OCR routine search FORMATS other than PDF? If yes, what
formats can the OCR search?
6. What are the resolution requirements for your OCR routines?
7. Can I change the GUI to a) add functionality or error checking
and b) to look personalized with CSS?
8. Can the search engine search using both requested metadata
element values and keywords from the document contents?
9. Can I start with keywords from the document contents and then
later filter the results using user inputted metadata element values?
10. Can I start with user input metadata element values and then
later filter down the results with document contents?
11. After an initial search, can I refine my search by only looking
at the results of the previous search?
Thanks,
Belinda
Re: JackRabbit Search Engine Questions
Posted by Marcel Reutegger <ma...@gmx.net>.
Hi Belinda,
Belinda Randolph wrote:
> 1. Can I replace the JackRabbit search engine with my own?
Yes you can, there are several interfaces you have to implement. See interface
QueryHandler for a starting point:
http://svn.apache.org/repos/asf/jackrabbit/tags/1.3/jackrabbit-core/src/main/java/org/apache/jackrabbit/core/query/QueryHandler.java
> 2. Does your search engine look through actual document contents - as a
> background process or at the time of the actual user search?
Whether text is extracted from documents and indexed when the document is saved
or deferred to a a number of background threads is configurable.
> 3. What FORMATs of actual documents does your search engine look at?
> (Ascii, Microsoft, PDF, etc.)
The currently supported formats are:
- Microsoft Word, Excel, PowerPoint
- PDF
- Open Office Documents (text, spreadsheet, presentation, etc.)
- RTF
- HTML
- XML
Text extraction in Jackrabbit is extensible. See:
http://svn.apache.org/repos/asf/jackrabbit/tags/1.3/jackrabbit-text-extractors/src/main/java/org/apache/jackrabbit/extractor/TextExtractor.java
> 4. When searching the contents of a PDF file, does the background
> process, using OCR, create an additional file in another format? What
> format?
The text extractor in Jackrabbit does not use OCR technology, but if you have an
existing java solution you may easily integrate it into Jackrabbit.
> 5. Does your OCR routine search FORMATS other than PDF? If yes, what
> formats can the OCR search?
n/a
> 6. What are the resolution requirements for your OCR routines?
n/a
> 7. Can I change the GUI to a) add functionality or error checking and
> b) to look personalized with CSS?
Jackrabbit is a content repository infrastructure and does not come with a user
interface. You may use any existing JCR compliant application on top of Jackrabbit.
> 8. Can the search engine search using both requested metadata element
> values and keywords from the document contents?
Yes, this is possible.
> 9. Can I start with keywords from the document contents and then later
> filter the results using user inputted metadata element values?
Yes, you would simply execute a second query that includes metadata values.
> 10. Can I start with user input metadata element values and then later
> filter down the results with document contents?
Yes, you would execute the first query with just the metadata values and then a
second one with additional keywords entered by the user.
> 11. After an initial search, can I refine my search by only looking at
> the results of the previous search?
Yes, you would simply execute the initial query again with additional search terms.
regards
marcel
Re: JackRabbit Search Engine Questions
Posted by Phillip Rhodes <sp...@rhoderunner.com>.
I am a newbie, but see my answers below
----- Original Message -----
From: "Belinda Randolph" <be...@jpl.nasa.gov>
To: users@jackrabbit.apache.org
Sent: Wednesday, May 2, 2007 4:42:23 PM (GMT-0500) America/New_York
Subject: JackRabbit Search Engine Questions
I am in the process of evaluating 10 repository solutions for my project.
I have several questions to ask in order to make my decisions.
1. Can I replace the JackRabbit search engine with my own?
yes you can, but why would you? I migrated to jackrabbit just so I could retire all my search code (written in lucene). Of course, you could write you own crawler/indexer to access the jackrabbit repository.
2. Does your search engine look through actual document contents -
as a background process or at the time of the actual user search?
The document is indexed when it is added to the repository. When the user searches, it is executing the search against a previouly built index. Very fast.
3. What FORMATs of actual documents does your search engine look
at? (Ascii, Microsoft, PDF, etc.)
All those formats, and more. You can easily create new ones if you like, you will have to set the mime type on the content that you add to get your custom indexer to run against it.
4. When searching the contents of a PDF file, does the background
process, using OCR, create an additional file in another format? What format?
When the pdf is added to jackrabbit, text is extracted from the pdf and added to the search index. No OCR involved. Just text extracted from the pdf. If the PDF contains only images, it will not do any ocr on those images.
5. Does your OCR routine search FORMATS other than PDF? If yes, what
formats can the OCR search?
There is no OCR technology involved, rather the text in the microsoft word document, etc. is extracted from the file using a library that understands the MS/PDF binary file format, so no OCR is necessary.
6. What are the resolution requirements for your OCR routines?
NO OCR involved with jackrabbit, but keep in mind, that we have the use of libraries that understand MS Word/PDF/Etc formats that can extract the textual content of the files.
7. Can I change the GUI to a) add functionality or error checking
and b) to look personalized with CSS?
Jackrabbit does not have a gui, so you are in total control of it. Some folks (like me) have application components written that allow easy creating of gui's to read/write/access jackrabbit content.
8. Can the search engine search using both requested metadata
element values and keywords from the document contents?
yes.
9. Can I start with keywords from the document contents and then
later filter the results using user inputted metadata element values?
yes.
10. Can I start with user input metadata element values and then
later filter down the results with document contents?
Yes.
11. After an initial search, can I refine my search by only looking
at the results of the previous search?
don't know.
Thanks,
Belinda