Searching in DB files - Any search engine available ??

Question

HELP:
I have several TIFF images I need to make it available on web as
searchable content.
Option 1:
I am planning of OCRing the images to text files. The file name
will be a unique ID. A DB table will contain meta data inforamtion
about the file. A search engine (??) can index these files. And the
results contain summary from the OCR text. And the resulting head
line will contain link to a JSP passing the UNIQUE ID as arguement
and JSP will read the tiff image for display. Thus I can track who
viewing the image and how many times.
Now any one know if there is a free search engine that can do this ?
I am looking into Lucene but I think it does not return summaries and
I am not sure if I can manipulate headlines..
Option 2: I can store the images and text in a DB (Oracle) is there
a search eingine that can index these ocred text files in db and
the result should be heading (from db meta data) linked to jsp with
doc id.
Example:
This is test <- from meta data
<p>This document contain result of your search... <- from ocr text

Answer 1

Lucene will allow you to do what you need -- it is content-agnostic. It has a concept of Document objects, which can contain one or more Field objects. You could have several Fields, like this:
Body -> contains the text of the file
Title -> contains the title
URL -> a web-friendly URL to retrieve the image?
PreviewURL -> a web-friendly URL to retrieve a preview?
Summary -> Short document summary
... and so on...
You can structure the Document object any way you want... Fields can be marked as stored, unstored (but still indexed...), etc.
You will need to write a couple of pieces of code to implement a lucene-based search engine:
1) Document indexer -- this is just a class that spiders through all your documents and adds them to the index. For your case, it would loop through the database and generate an index entry for each part. This should probably be on a timer to keep the index up-to-date.
2) Search UI -- this just presents the user with some way to search the documents.
Anyway, the devil's in the details, and this is no small project. Hope this gives you a good start.