|
Re: query from more than one search engine
by MZSanford (Curate) on Nov 26, 2001 at 20:39 UTC
|
We are not a code base, we do not (usually) write code for people. As for the multi search engine problem, would could look into the various WWW::Search:: modules, and write something to hold them together ... or, you could find an existing search engine that searches others, and query only that .. that may work too.
But, just asking us to write your homework (as this is what it sounds like), without posting anything, is a bit of a strech. Hopefully the WWW::Search:: modules have what you need, and you will post any problems you have during the coding.
i had a memory leak once, and it ruined my favorite shirt. | [reply] |
Re: query from more than one search engine
by frankus (Priest) on Nov 26, 2001 at 20:42 UTC
|
Lego coder post ;) WWW::Search on CPAN?
Perl is about the best language to do this in, PHP might also do,
it's going to be a tough task stripping meaningful data from web-pages unless you're familiar with that.
But CPAN has some excellent resources and they're are a lot of good minds here.
There is an art to posting questions, which sometimes provides answers by itself:
Phrase your question well and read it back, see if your needs are immediately clear.
--
Brother Frankus.
¤ | [reply] |
Re: query from more than one search engine
by Hero Zzyzzx (Curate) on Nov 26, 2001 at 20:41 UTC
|
Doubt you'll get anyone to do your work for you. PerlMonks are charitable, but not foolish. . .
Post what you've got and people will help you work through it.
Buy O'Reilly's "Learning Perl", read it, love it, understand it. Then buy O'Reilly's "CGI Programming with Perl" and do the same.
Check out the CPAN for code- Someone has probably already done some of your work for you.
Good luck. People will help you learn, but won't do it for you (unless you pay them, I suppose).
-Any sufficiently advanced technology is
indistinguishable from doubletalk.
| [reply] |
Re: query from more than one search engine
by fuzzysteve (Beadle) on Nov 26, 2001 at 20:47 UTC
|
I won't write it for you (don't have the time, sorry), but the module you should look at is LWP::UserAgent It will allow you to emulate a browser uerying the search engine. You'll have to build up the query string with the requests, but in most cases thats just a case of inserting the search strings into the query string.
Google for example, uses query strings like
http://www.google.com/search?q=blam
to search for 'blam'
Then you'll need to parse your way through the returned page and pull the links to outside pages. Harder, but not too bad. Google for example, looks liek if you pull two lines from each <p> tag you'll get the lines with the details you'll need.
The hardest bit, to my mind? Merging them in a sensible way. what counts as sensible? | [reply]
[d/l] |
|
|
that is what i do.. i put the the strings like the one up there. but, i could only put 1 string/address..not 3...that is my problem....how do i put 3 strings/addresses? ..or how do i make the query passes at least 3 search engine before giving the results?
| [reply] |
|
|
you'll have to grab all three (seperate calls), parse out the links, then compile a page from those links.
pseudo code:
page1=grab link1 into variable;
page2=grab link2 into variable;
page3=grab link3 into variable;
define an array of hashs;
parse links from page1 pushing into the array;
parse links from page2 pushing into the array;
parse links from page3 pushing into the array;
weight the array in a sensible fashion;
output page using the array;
| [reply]
[d/l] |
|
|
|
|
|