Greetings!
My Java app, using SolrJ, now successfully does searches. I've used the
web interface to do a full-text indexing and for each new entry added
through my app, I have it add to this index.
But now I want to use SolrJ to also do spell checking. I have read
several documents on this and examined a couple of Java examples, but
one main question still persists. Let's first assume that I have my
configuration XML files set up correctly and I can spell-check a word
through the web interface using something like
.../spell?q=missspelled&spellcheck=on. Assume also that the end user
has typed in a paragraph and is about to submit the text. In the
current implementation of my software, using SolorJ API, the text will
get parsed into words and the words will be added to the search index.
For spell-checking, however; I am puzzled.
Is it up to me, the programmer, to parse the text into individual words
and determine which words are misspelled, then run the query on a
misspelled word to get a list of suggestions for that misspelled word??
Or does Solr itself parse the text string into words and run a query on
every word, thus indicating which words are misspelled by the non-zero
list of suggestions? Or is there a third option I haven't thought of
(like, spell-check as I type)??
I'm just trying to picture the behavior in my head so I know what
programming approach to take. Thanks for the help!
Mark
