The document seems to point to using AutoPhrasingTokenFilter, putting an underscore to the multi-term or changing to index time synonyms.
I'm also thinking of putting the synonyms onto a database or query some thesaurus website when the using enter the search key, instead of using the SynonymFilterFactory. For this, once user enter a search key, the program will retrieve the list of synonyms. Then I'll append the list to the search parameters (ie: q). I'll use the boosting relevancy to give the original term a higher boost, and the synonyms a lower boost. Is this a good solution? Regards, Edwin On 8 May 2015 17:40, "Alessandro Benedetti" <benedetti.ale...@gmail.com> wrote: > I found this very interesting article that I think can help in better > understanding the problem : > > http://lucidworks.com/blog/solution-for-multi-term-synonyms-in-lucenesolr-using-the-auto-phrasing-tokenfilter/ > > And this : > > http://opensourceconnections.com/blog/2013/10/27/why-is-multi-term-synonyms-so-hard-in-solr/ > > Take a look and let me know ! > > 2015-05-08 10:26 GMT+01:00 Zheng Lin Edwin Yeo <edwinye...@gmail.com>: > > > Thanks for explaining the information. > > > > Currently I'm only using the comma-separated list of words and only using > > the synonym filter at query time. I find that when I set expend = true, > > there's quite a number of irrelevant results that came back, and this > > didn't happen when I set expend = false. > > > > I've yet to try the lists of words with the symbol "=>" between them. I'm > > trying to solve the multi-word synonyms too, and I found that enclosing > the > > multi-word in quotes will solve the issue. But this creates problem and > the > > original token is not return if I enclose single word in quotes. > > > > Will using the lists of words with the symbol "=>" between them better > than > > the comma-separated list of words to cater to the multi-word synonyms? > > > > Regards, > > Edwin > > > > > > > > On 8 May 2015 at 17:10, Alessandro Benedetti <benedetti.ale...@gmail.com > > > > wrote: > > > > > Let's explain little bit better here : > > > First of all, the SynonimFilter is a Token Filter, and being a Token > > Filter > > > it can be part of an Analysis pipeline at Indexing and Query Time. > > > As the different type of analysis explicitly explains when the > filtering > > > happens, let's go to the details of the synonyms.txt. > > > This file contains a set of lines, each of them describing a synonym > > > policy. > > > There are 2 different syntaxes accepted : > > > > > > > > > > > > > > > *couch,sofa,divanteh => thehuge,ginormous,humungous => largesmall => > > > tiny,teeny,weeny* > > > > > > > > > - A comma-separated list of words. If the token matches any of the > > > words, then all the words in the list are substituted, which will > > > include > > > the original token. > > > > > > > > > - Two comma-separated lists of words with the symbol "=>" between > > them. > > > If the token matches any word on the left, then the list on the > right > > is > > > substituted. The original token will not be included unless it is > also > > > in > > > the list on the right. > > > > > > > > > Related the "expand" param, directly from the official Solr > > documentation : > > > > > > expand: (optional; default: true) If true, a synonym will be expanded > to > > > all equivalent synonyms. If false, all equivalent synonyms will be > > reduced > > > to the first in the list. > > > > > > So, starting from this definition let's answer to your questions: > > > > > > 1) Related the expand the definition seems quite clear, if anything > > strange > > > is occurring to you, let me know > > > 2) Related your second question, it depends on your synonym.txt file, > if > > > you are not using the => syntax, you are going to always retrieve all > > > the synonyms( > > > included the original term) > > > > > > If you need more info let me know, it can strictly depends how you are > > > using the filter as well ( indexing ? querying ? both ? ) > > > Example : > > > If you are using the filter only at Indexing time, then using the => > > syntax > > > will prevent the user to search for the original token in the > synonym.txt > > > relation. > > > Because it will not appear in the index. > > > > > > Cheers > > > > > > > > > 2015-05-08 9:24 GMT+01:00 Zheng Lin Edwin Yeo <edwinye...@gmail.com>: > > > > > > > Hi, > > > > > > > > Will like to check, for the SynonymFilterFactory, I have the > following > > in > > > > my synonyms.txt: > > > > > > > > Titanium Dioxides, titanium oxide, pigment > > > > pigment, colour, colouring material > > > > > > > > If I set expend=false, and I search for q=pigment, I will get results > > > that > > > > matches pigment, Titanium Dioxides and titanium oxide. But it will > not > > > > maches colour and colouring materials, as all equivalent synonyms > will > > > only > > > > matches those first in the list. > > > > > > > > If I set expend=false, and I search for q=pigment, I'll get results > > that > > > > matches everything in the list (ie: Titanium Dioxides, titanium > oxide, > > > > colour, colouring material) > > > > > > > > Is my understand correct? > > > > > > > > Also, I will like to check, how come if I search q="pigment" > (enclosed > > in > > > > quotes), I only get matches for Titanium Dioxides and not pigment? > > > > > > > > Regards, > > > > Edwin > > > > > > > > > > > > > > > > -- > > > -------------------------- > > > > > > Benedetti Alessandro > > > Visiting card : http://about.me/alessandro_benedetti > > > > > > "Tyger, tyger burning bright > > > In the forests of the night, > > > What immortal hand or eye > > > Could frame thy fearful symmetry?" > > > > > > William Blake - Songs of Experience -1794 England > > > > > > > > > -- > -------------------------- > > Benedetti Alessandro > Visiting card : http://about.me/alessandro_benedetti > > "Tyger, tyger burning bright > In the forests of the night, > What immortal hand or eye > Could frame thy fearful symmetry?" > > William Blake - Songs of Experience -1794 England >
