So you've got a dynamic site, filled with all sorts of user inputs, whether
it be a 'phorum', or like
my own site at knowpost.com. The site htdig.org will take care of indexing and searching your html pages, but if you are like me, you have very few html pages, and must of your
"content" resides in BLOBs in your database. You can't do anything useful using a like %searchword% query, it just isn't coming back relevant.
There has to be a better way, and indeed there is, with a few easy steps.
Here's how to slap one together:
Part one: BNR--Blob Noise Reduction
The first problem with your content is that it is filled with clunky
"noisewords," like "a,the,where,look"
Things that are there to help us humans to communicate, but really don't
have anything to do with relevance.
We gotta get rid of those. I've included a big list of noisewords
(noisewords.txt) for you to use, modify
or mutilate. Essentially, what we're trying to do here is get all those
noisewords out of your data, and build
a table with two columns, the word, and its indicator (the content
associated with it). We want something that
will eventually look like this:
+------+------------+
| qid | word |
+------+------------+
| 6 | links |
| 5 | Fire |
| 5 | topics |
| 5 | related |
| 5 | Shakespeare|
| 4 | people |
| 4 | Knowpost |
| 3 | cuba |
| 3 | cigar |
+------+------------+
Lets create our table now--
CREATE TABLE search_table(
word VARCHAR(50),
qid INT)
Next, since you want to make all your data compatible, not just new data, we
need to grab your sticky
blobs, and their identifiers out of your database:
<?php
$query = "SELECT blob,identifier FROM your_table";
$result = mysql_query($query);
$number = mysql_numrows($result);
$j = 0;
WHILE ($j < $number) {
/* Your "blob" */
$body = mysql_result($result,$j,"blob");
/* Your "identifier */
$qid = mysql_result($result,$j,"qid");
/* Open the noise words into an array */
$noise_words = file("noisewords.txt");
$filtered = $body;
/* Got to put a space before the first word in the
body, so that we can
recognize the word later
*/
$filtered = ereg_replace("^"," ",$filtered);
/* Now we suck out all the noisewords, and transform
whats left into an array
*/
/* Brought to you by poor ereg coding! */
for ($i=0; $i < count($noise_words); $i++) {
$filterword = trim($noise_words[$i]);
$filtered =
eregi_replace(" $filterword "," ",$filtered);
}
$filtered = trim($filtered);
$filtered = addslashes($filtered);
$querywords = ereg_replace(",","",$filtered);
$querywords = ereg_replace(" ",",",$querywords);
$querywords = ereg_replace("\?","",$querywords);
$querywords = ereg_replace("\(","",$querywords);
$querywords = ereg_replace("\)","",$querywords);
$querywords = ereg_replace("\.","",$querywords);
$querywords = ereg_replace(",","','",$querywords);
$querywords = ereg_replace("^","'",$querywords);
$querywords = ereg_replace("$","'",$querywords);
/* We should now have something that looks like
'Word1','Word2','Word3' so lets turn it into an array
*/
$eachword = explode(",", $querywords);
/* and finally lets go through the array, and place each
word into the database, along with its identifier
*/
for ($k=0; $k < count($eachword); $k++) {
$inputword = "INSERT INTO search_table
VALUES($eachword[$k],$qid)";
mysql_query($inputword);
}
/* Get the next set of data */
$j++;
}
?>
That script just handles your old data. You'll want to include a similar
function to strip the noisewords
out for every time new information comes into your database, through user
input, your input, etc...
so that your search engine is updated on the fly.
Part 2: Searching the Table
Now you have an easy to-use table of keywords and their associations. How do
you query this table? Here's
what I do:
First I format each searchterms passed into the script as
'word1','word2','word3' and stick it in a string
called
$querywords.
Then I throw them into this SQL query:
SELECT count(search_table.word) as score, search_table.qid,your_table.blob
FROM search_table,your_table
WHERE your_table.qid = search_table.qid AND search_table.word
IN($querywords)
GROUP BY search_table.qid
ORDER BY score DESC";
Set that query to
$search, and print out the results like so:
<?php
$getresults = mysql_query($search);
$resultsnumber = mysql_numrows($getresults);
IF ($resultsnumber == 0) {
PRINT "Your search returned no results. Try other keyword(s).";
} ELSEIF ($resultsnumber > 0) {
PRINT "Your search returned $resultsnumber results<BR>Listing
them in order of relevance<BR><BR>";
for($count = 0; $count < $resultsnumber; $count++) {
$body = mysql_result($getresults,$count,"blob");
$qid = mysql_result($getresults,$count,"qid");
//tighten up the results
$body2print = substr($body, 0, 100);
$cnote = $count+1;
PRINT "$cnote. <a href=yourcontent.php3?qid=$qid>
<i>$body2print...</i></a><BR>";
}
}
?>
Presto, you've got keyword searching for your database, complete with
relevancy ranking. It may not be Google
or altavista.
It may not support all those fancy boolean operators, or excite's (*cough*)
conceptual mapping technology. But
it works, its quick
and enough to handle your user's demand.
Here's that list of noisewords:
noisewords.txt
--------------
a
about
after
ago
all
almost
along
also
am
an
and
answer
any
anybody
anywhere
are
aren't
around
as
ask
at
bad
be
been
before
being
best
better
between
big
but
by
can
can't
come
could
couldn't
day
did
didn't
do
does
don't
down
each
either
else
even
ever
every
everybody
everyone
far
find
for
found
from
get
go
going
gone
good
got
had
has
have
haven't
having
her
here
hers
him
his
home
how
href
I
if
in
into
is
isn't
it
its
know
large
less
like
little
looking
look
many
me
more
most
must
my
near
never
new
news
no
none
not
nothing
of
off
often
old
on
once
only
or
other
our
ours
out
over
page
please
question
rather
recent
she
should
sites
small
so
some
something
sometime
somewhere
than
true
thank
that
the
their
theirs
them
then
there
these
they
this
those
though
through
thus
time
times
to
too
under
until
untrue
up
upon
use
users
version
very
via
want
was
way
web
were
what
when
where
which
who
whom
whose
why
wide
will
with
within
without
world
worse
worst
would
www
yes
yet
you
your
yours
How
--clay