filters - how to edit ignore word(s) ?

Help
wyat
2012-02-26
2012-08-29
  • wyat

    wyat - 2012-02-26

    hey - im using rssowl to search craigslist for an apartment-- when i try to
    set a filter to find the phrase "no smoking" rssowl ignores the "no" so it
    finds instead "smoking", which could mean alot of different things. i would
    like to remove "no" from the ignore words so i can include it in an exact
    phrase filter- how can i do this? it will save me alot of time

    thank you for help!

     
  • Anonymous

    Anonymous - 2012-02-26

    Well you can just say "Entire News" + "doesn't contain" + "apartment". Check
    the dropdowns.

     
  • Anonymous

    Anonymous - 2012-02-26

    Eh I meant "doesnt contain smoking" :). Good luck.

     
  • wyat

    wyat - 2012-02-26

    bpasero, thanks for the reply-

    I can't do it that way, because sometimes the ads say "smoking ok outside" or
    "smoking permitted in back" so to exclude "smoking" entirely would cut out
    alot of possible places for me... I need to be able to filter with the phrase
    "no smoking" or "no smokers"

    This must be possible somehow? Maybe the ignore words are in an xml file or
    something hehe? I'm not a programmer ;(

    thanks again!

     
  • genodeftest

    genodeftest - 2012-02-26

    Sorry, but this can not be done with RSSOwl or any „normal“ software. What you
    want is a parsing feature for human language. There is no software on earth
    having it fully implemented…

    Maybe you're lucky and find some less complicated (thus limited) software –
    awfully expensive.

     
  • wyat

    wyat - 2012-02-27

    hehe- all i want to do is be able to filter the exact phrase "no smoking" but
    rssowl ignores the word "no" so i wanted to remove it from the ignore words,
    cant be that hard ;)

     
  • genodeftest

    genodeftest - 2012-02-27

    No, but you will get problems with "smoking permitted" and "smoking not
    permitted", "smoking ok outside", "smoking unwelcome", "smoking ban" "no-
    smoking rule", …

     
  • airfern

    airfern - 2012-03-28

    You probably need to use AND logic for that kind of filter. Try to use plus
    symbol: no+smoking

     
  • CallMeAndy

    CallMeAndy - 2012-04-11

    "Normal" programs intending at searchin and filtering large amounts of data
    would use a database to store downloaded feeds. Every piece of software on
    earth utilising any SQL supporting database would find this a walk in the
    park.

    And er there is rather a lot, and er you dont need any kind of special parsing
    feature just bog standard SQL -

    I see Feedreader has a MYSQL version in beta. That might be worth a look if
    you want to make any sort of easy to write but powerful queries. Though I
    havent tried it myself so cant comment on its strengths or weaknesses.

     
  • genodeftest

    genodeftest - 2012-04-12

    RSSOwl uses db4o for database and Apache Lucene for Indexing/Search, which is
    one of the most powerful search engines. And db4o is not (primarily) using
    String-based queries (like SQL), since that would eat too much CPU. For
    special search queries see http://wiki.apache.org/lucene-
    java/LuceneFAQ#Searching
    .

     
  • CallMeAndy

    CallMeAndy - 2012-04-12

    Fair comment - I am not going to tell you guys how to build your software I
    dont presume to be qualified but the element of the statement at no 5 that is
    not reasonable is:

    "this can not be done with RSSOwl or any „normal“ software

    maybe it cant be done with RSSOwl but how can software using db4o be
    considered as using a normal backend in relation to normal software, if it
    susggested that databases using SQL are not normal. The use of db4o is an
    architectural choice but its not appropriate to diminish other architectures
    particularly those using far more accepted solutions.

    I think you as developers however, all be it you maybe making perfectly
    legitimate choices for other reasons, are missing an opportunity. SQL use on
    the client side against locally poulated datbases would be a huge draw on RSS
    reader methods. Obviously the guys at Feedreader would agree with me because
    they're SQL solution is at a fee whereas their Free version does not employ
    SQL.

    The point of RSS is to reduce the information overload or at least the ability
    to keep up with the important bits. Being able to query the database with an
    efficient query mechanism is far more important than local CPU time.

    A couple of seconds longer on a large database or having to look through
    potentially 1000's of feeds of no value to me. Which is best for the user?

     
  • genodeftest

    genodeftest - 2012-04-12

    Hey,

    1. I am no RSSOwl developer (bpasero does the work, but he could need some help)
    2. against SQL:
      • it is slow (you're right, this is no big deal on client side)

    • compared to SQL-based RDBMS db4o (object-oriented DBMS) is easier to handle
    and write in a Object-Oriented programming language like Java. This is no
    matter of usability but of effort writing code.

    • maybe db4o is integrated into lucene or the other way round, maybe some
    other reasons bpasero choose it

    1. I don't know whether RSSOwl/Lucene does use some SQL functionality of db4o or not.
     
  • CallMeAndy

    CallMeAndy - 2012-04-12

    I would love to help but I need to make for myself an income first. Going down
    the pan here.

    I dont see why you need an object oriented database. your layout is perfectly
    suited to a relational model, there is nothing you are trying to do in the
    reader that warrents an object oriented model.

    Easier to write...

    The first example on the first googled return - how much more difficult can it
    be. I have used OO lnaguages with SQL forever - there is nothing difficult in
    it.

    import java.sql.Connection;  
     import java.sql.DriverManager;  
     import java.sql.ResultSet;  
     import java.sql.Statement;  
     public class ConnectSQLite 
     {  
      public static void main(String[] args) 
      {  
         Connection connection = null;  
         ResultSet resultSet = null;  
         Statement statement = null;
    
         try 
         {  
             Class.forName("org.sqlite.JDBC");  
             connection = DriverManager.getConnection("jdbc:sqlite:D:\\testdb.db");  
             statement = connection.createStatement();  
             resultSet = statement  
                     .executeQuery("SELECT EMPNAME FROM EMPLOYEEDETAILS");  
             while (resultSet.next()) 
             {  
                 System.out.println("EMPLOYEE NAME:"  
                         + resultSet.getString("EMPNAME"));  
             }  
         }
    

    SQL itself is a breeze for simple SELECT and UPDATE statements.

    Why do you need Lucerne? This is a genuine question. Surely there are many
    other ways to crawl and of course Lucerne maybe recognized as a first class
    text indexer but most rdbms will do it adequatly and like I said I dont mind
    waiting a couple of mins extra if need be to save myself hours of work.

    At the end of the day I am not going to tell you guys how to design your
    software, and I am sure there are a lot of aspects to your architectural
    decisions than the immediatly discussed ones. Also of course, the headache of
    changing course may in itself seem expensive.

    But consider this - I might be a typical dissalusioned potential user of
    RSSOwl (in fact of all RSS readers I have tried - and that is a lot over 6 or
    7 maybe more years) but I keep trying every now and then because the concept
    of a news aggregator that can find exactly what I want, on a daily basis,
    should save me hours of work on a daily basis. Instead I am still in the same
    place, years on, dissalusioned with the various readers abilities to reduce my
    information swamp to a puddle, although in fact this time a rather improved
    position because it has led me to google reader.

    This should set alarm bells ringing for you guys as the days of desktop
    readers might seriously be numbered. The user interface is not the SQL I have
    been looking for but its nearly as damn good and it maybe better for what I
    want. There are no semantics to learn for anybody working for me. I can set up
    different queries with different links, and I can embed them on a single local
    html document with the simplest of script so that I can scan all queries at
    once ....

    Ironically this round of checking out the RSS readers I had hoped would come
    to maturity has led me to use RSSOwl; sadly I am using it to collect feeds
    that I am subscribing to and querying in google reader.

     

Log in to post a comment.