I would like to know if Solr suggestions improve with every search attempt. If not is there a method to do so. I tried configuring Solr suggester but doesn't seem to work.
Can anyone point me in the right direction. Ive tried everything I could find out, so far I'm not making any progress.
Examples say setting the Solr config and all did as they had mentioned ( in Solr Wiki ) but still couldn't get Suggest to work. Im using Solr 4.7, is this a version issu?
Here is the fields that I attempt to use for suggestion
<field name="suggestions" type="suggest_text" indexed="true" stored="true" />
<fieldType name="suggest_text" class="solr.TextField">
<analyzer>
<tokenizer class="solr.StandardTokenizerFactory"/>
<filter class="solr.WordDelimiterFilterFactory" generateWordParts="1" generateNumberParts="1" catenateWords="0" catenateNumbers="0" catenateAll="0" splitOnCaseChange="1"/>
<filter class="solr.RemoveDuplicatesTokenFilterFactory" />
<filter class="solr.LowerCaseFilterFactory"/>
<filter class="solr.SynonymFilterFactory" synonyms="hitad_synonyms.txt" ignoreCase="true" expand="true"/>
<filter class="solr.EdgeNGramFilterFactory" maxGramSize="20" minGramSize="1"/>
</analyzer>
</fieldType>
Can someone shed some light on this issue?
Related
I've got this fieldType on my Solr implementation
<fieldType name="suggestion_text" class="solr.TextField" positionIncrementGap="100">
<analyzer type="index">
<tokenizer class="solr.WhitespaceTokenizerFactory"/>
<filter class="solr.WordDelimiterFilterFactory"
generateWordParts="1"
generateNumberParts="1"
splitOnNumerics="1"
preserveOriginal="1"
/>
<filter class="solr.EdgeNGramFilterFactory" maxGramSize="100"/>
<filter class="solr.LowerCaseFilterFactory"/>
</analyzer>
<analyzer type="query">
<tokenizer class="solr.StandardTokenizerFactory"/>
<filter class="solr.LowerCaseFilterFactory"/>
</analyzer>
</fieldType>
This works fine for almost every model I've got. For example for models AB1234, I can search 1234 and it finds it. But there's a particular case that I want to include and I'm trying to find a better solution than the current one:
Let's say AB is the manufacturer and 1234 is the actual part number, but in my database they are saved as AB1234. It I've got an A0 manufacturer, and A01234 partnumber, with the current implementation if i search 1234 i wont find it.
I found a workaround transforming the EdgeNGramFilterFactory into a NGramFilterFactory, but that's not the solution I want. I want the Solr to be able to search excluding the first two characters if they are letter+number or in the extreme case, but I need it to search with A0 and without A0.
I don't know if I was clear. Anyway i tryied with regular expressions, creating a new field and using this filter on it:
<filter class="solr.PatternReplaceFilterFactory" pattern="(A0)" replacement="" replace="all" />
or
<filter class="solr.PatternReplaceFilterFactory" pattern="[a-zA-Z][0-9]" replacement="" replace="all" />
but this is not giving expected results.
Can you help me? Thank you
Please try the below field type for your problem.
<fieldType name="text_en_splitting_test" class="solr.TextField" positionIncrementGap="100" autoGeneratePhraseQueries="true">
<analyzer type="index">
<tokenizer class="solr.WhitespaceTokenizerFactory"/>
<filter class="solr.WordDelimiterGraphFilterFactory" generateWordParts="1" generateNumberParts="1" catenateWords="1" catenateNumbers="1" splitOnNumerics="1" preserveOriginal="1"/>
<filter class="solr.LowerCaseFilterFactory"/>
</analyzer>
<analyzer type="query">
<tokenizer class="solr.KeywordTokenizerFactory"/>
<filter class="solr.LowerCaseFilterFactory"/>
</analyzer>
</fieldType>
This worked for below search terms.
AB1234
AB
1234
Please find the screenshot of the solr analysis page for the suggested field type with the search terms.
I've found an interesting query for Solr and it returns search results, but I don't understand, what is the purpose of slash symbol between the words?
duties:health/nurse
Anybody knows? Please, help.
Simple. You can look at the analyzer chain to understand what happens.
My guess is that the analyzer chain turns the / into a space - which makes the query into
duties: health nurse
To find out your analyzer chain from the configuration - start by checking the type of the field
For example
<field name="health" type="text_general" indexed="true" stored="true" required="true"/>
Now we look for the definition of the type
<fieldType name="text_general" class="solr.TextField" positionIncrementGap="100">
<analyzer type="index">
<tokenizer class="solr.StandardTokenizerFactory"/>
<filter class="solr.StopFilterFactory" ignoreCase="true" words="stopwords.txt" />
<!-- in this example, we will only use synonyms at query time
<filter class="solr.SynonymFilterFactory" synonyms="index_synonyms.txt" ignoreCase="true" expand="false"/>
-->
<filter class="solr.LowerCaseFilterFactory"/>
</analyzer>
<analyzer type="query">
<tokenizer class="solr.StandardTokenizerFactory"/>
<filter class="solr.StopFilterFactory" ignoreCase="true" words="stopwords.txt" />
<filter class="solr.SynonymFilterFactory" synonyms="synonyms.txt" ignoreCase="true" expand="true"/>
<filter class="solr.LowerCaseFilterFactory"/>
</analyzer>
</fieldType>
As you can see, we have an index analyzer and a query analyzer.
My query analyzer would turn / in the query into something else by using the StandardTokenizerFactory.
From the solr wiki:
solr.StandardTokenizerFactory
A good general purpose tokenizer that strips many extraneous characters and sets token types to meaningful values. Token types are only useful for subsequent token filters that are type-aware of the same token types. There aren't any filters that use StandardTokenizer's types.
I am thinking that health/nurse is being viewed as a string literal as there are no spaces between. Health / nurse should yield different results than health/nurse, correct? If so, then health/nurse must be an indexed term in your documents.
I'm using Solr to index products on a clothing website. At the moment I'm trying to get Solr to match t-shirt based on the search term tshirt, but I'm a little bit lost as to what filters to I need.
This is the general purpose field type that I'm using to index most fields at the moment:
<fieldType name="text_en" class="solr.TextField" positionIncrementGap="100">
<analyzer type="index">
<tokenizer class="solr.StandardTokenizerFactory"/>
<filter class="solr.StopFilterFactory" ignoreCase="true" words="lang/stopwords_en.txt" enablePositionIncrements="true" />
<filter class="solr.LowerCaseFilterFactory"/>
<filter class="solr.EnglishPossessiveFilterFactory"/>
<filter class="solr.KeywordMarkerFilterFactory" protected="protwords.txt"/>
<filter class="solr.PorterStemFilterFactory"/>
<filter class="solr.WordDelimiterFilterFactory" catenateWords="1" catenateAll="1" preserveOriginal="1"/>
</analyzer>
<analyzer type="query">
<tokenizer class="solr.StandardTokenizerFactory"/>
<filter class="solr.SynonymFilterFactory" synonyms="synonyms.txt" ignoreCase="true" expand="true"/>
<filter class="solr.StopFilterFactory" ignoreCase="true" words="lang/stopwords_en.txt" enablePositionIncrements="true" />
<filter class="solr.LowerCaseFilterFactory"/>
<filter class="solr.EnglishPossessiveFilterFactory"/>
<filter class="solr.KeywordMarkerFilterFactory" protected="protwords.txt"/>
<filter class="solr.PorterStemFilterFactory"/>
<filter class="solr.WordDelimiterFilterFactory" catenateWords="1" catenateAll="1" preserveOriginal="1"/>
</analyzer>
</fieldType>
I tried removing WordDelimiterFilterFactory from the index and query analyzers, but it didn't help. Any advice/best practice would be really appreciated.
You'll want to have the WordDelimiterFilter further up your chain, and you'll want to use the Whitespace Tokenizer instead. The example on the wiki does just that.
The issue now is that the tokens are split into separate tokens earlier on, and the worddelimeterfilter is only seeing each token by itself. So it sees the t, then shirt and doesn't really have anything to do.
By using the whitespace tokenizer you'll get the WDF to see "t-shirt", allowing it to generate t, shirt, tshirt, etc.
Use the "Analysis" page under the Solr Admin to see each step in the analysis and what the result is.
I have a problem with a search with special characters in solr.
My document has a field "title" and sometimes it can be like "Titanic - 1999" (it has the character "-").
When i try to search in solr with "-" i receive a 400 error. I've tried to escape the character, so I tried something like "-" and "\-". With that changes solr doesn't response me with an error, but it returns 0 results.
How can i search in the solr admin with that special character(something like "-" or "'"???
Regards
UPDATE
Here you can see my current solr scheme https://gist.github.com/cpalomaresbazuca/6269375
My search is to the field "Title".
excerpt from the schema.xml:
...
<!-- A general text field that has reasonable, generic
cross-language defaults: it tokenizes with StandardTokenizer,
removes stop words from case-insensitive "stopwords.txt"
(empty by default), and down cases. At query time only, it
also applies synonyms. -->
<fieldType name="text_general" class="solr.TextField" positionIncrementGap="100">
<analyzer type="index">
<tokenizer class="solr.StandardTokenizerFactory"/>
<filter class="solr.StopFilterFactory" ignoreCase="true" words="stopwords.txt" enablePositionIncrements="true" />
<!-- in this example, we will only use synonyms at query time
<filter class="solr.SynonymFilterFactory" synonyms="index_synonyms.txt" ignoreCase="true" expand="false"/>
-->
<filter class="solr.LowerCaseFilterFactory"/>
</analyzer>
<analyzer type="query">
<tokenizer class="solr.StandardTokenizerFactory"/>
<filter class="solr.StopFilterFactory" ignoreCase="true" words="stopwords.txt" enablePositionIncrements="true" />
<filter class="solr.SynonymFilterFactory" synonyms="synonyms.txt" ignoreCase="true" expand="true"/>
<filter class="solr.LowerCaseFilterFactory"/>
</analyzer>
</fieldType>
...
<field name="Title" type="text_general" indexed="true" stored="true"/>
You are using the standard text_general field for the title attribute. This might not be a good choice. text_general is meant to be for huge chunks of text (or at least sentences) and not so much for exact matching of names or titles.
The problem here is that text_general uses the StandardTokenizerFactory.
<fieldType name="text_general" class="solr.TextField" positionIncrementGap="100">
<analyzer type="index">
<tokenizer class="solr.StandardTokenizerFactory"/>
<filter class="solr.StopFilterFactory" ignoreCase="true" words="stopwords.txt" enablePositionIncrements="true" />
<!-- in this example, we will only use synonyms at query time
<filter class="solr.SynonymFilterFactory" synonyms="index_synonyms.txt" ignoreCase="true" expand="false"/>
-->
<filter class="solr.LowerCaseFilterFactory"/>
</analyzer>
<analyzer type="query">
<tokenizer class="solr.StandardTokenizerFactory"/>
<filter class="solr.StopFilterFactory" ignoreCase="true" words="stopwords.txt" enablePositionIncrements="true" />
<filter class="solr.SynonymFilterFactory" synonyms="synonyms.txt" ignoreCase="true" expand="true"/>
<filter class="solr.LowerCaseFilterFactory"/>
</analyzer>
</fieldType>
StandardTokenizerFactory does the following:
A good general purpose tokenizer that strips many extraneous
characters and sets token types to meaningful values. Token types are
only useful for subsequent token filters that are type-aware of the
same token types.
This means the '-' character will be completely ignored and be used to tokenize the String.
"kong-fu" will be represented as "kong" and "fu". The '-' disappears.
This does also explain why select?q=title:\- won't work here.
Choose a better fitting field type:
Instead of the StandardTokenizerFactory you could use the solr.WhitespaceTokenizerFactory, that only splits on whitespace for exact matching of words. So making your own field type for the title attribute would be a solution.
Solr also has a fieldtype called text_ws. Depending on your requirements this might be enough.
To search for your exact phrase put inverted commas round it:
select?q=title:"Titanic - 1999"
If you just want to search for that special character then you will need to escape it:
select?q=title:\-
Also check:
Special characters (-&+, etc) not working in SOLR Query
If you know exactly which special characters you dont want to use then you can add this to the regex-normalize.xml
<regex>
<pattern>-</pattern>
<substitution>%2D</substitution>
</regex>
This will replace all "-" with %2D, so when you search, as long as you search for %2D instead of the "-" it will work fine
I spent a lot of time getting this done. Here is a clear step-by-step things to be done to query special characters in SolR. Hope it helps someone.
Edit the schema.xml file and find the solr.TextField that you are
using.
Under both, "index" and query" analyzers modify the
WordDelimiterFilterFactory and add types="characters.txt" Something like:
<fieldType name="text_general" class="solr.TextField" positionIncrementGap="100" multiValued="true">
<analyzer type="index">
<tokenizer class="solr.WhitespaceTokenizerFactory"/>
<filter catenateAll="0" catenateNumbers="0" catenateWords="0" class="solr.WordDelimiterFilterFactory" generateNumberParts="1" generateWordParts="1" splitOnCaseChange="1" types="characters.txt"/>
</analyzer>
<analyzer type="query">
<tokenizer class="solr.WhitespaceTokenizerFactory"/>
<filter catenateAll="0" catenateNumbers="0" catenateWords="0" class="solr.WordDelimiterFilterFactory" generateNumberParts="1" generateWordParts="1" splitOnCaseChange="1" types="characters.txt"/>
</analyzer>
</fieldType>
Ensure that you use WhitespaceTokenizerFactory as the tokenizer as
shown above.
Your characters.txt file can have entries like-
\# => ALPHA
# => ALPHA
\u0023 => ALPHA
ie:- pointing to ALPHA only.
Clear the data, re-index and query for the entered characters. It
will work.
I want to use synonym search in solr for multi keyword.
But It doesn't work correct.
I set the synonym "multi term" for "multerm" in synonym.txt. And I expect that Solr makes query-phrase for "multerm" just like "field:"multi term"~0 but "field:multi | field:term". So It can't do intimacy search for multi term synonym.
Any one has the best way for multi term synonym search in Solr? Help me please~
Here is how I handle multi-word synonyms. In my schema.xml, fieldType definition looks like:
<filter class="solr.SynonymFilterFactory" synonyms="synonyms.txt" ignoreCase="true" expand="true" tokenizer="solr.KeywordTokenizerFactory"/>
<fieldType name="custom_text_general" class="solr.TextField" positionIncrementGap="100">
<analyzer type="index">
<tokenizer class="solr.StandardTokenizerFactory"/>
<filter class="solr.StopFilterFactory" ignoreCase="true" words="stopwords.txt" enablePositionIncrements="true" />
<!-- We will use synonyms only at index time to keep querying fast-->
<filter class="solr.SynonymFilterFactory" synonyms="synonyms.txt" ignoreCase="true" expand="true" tokenizer="solr.KeywordTokenizerFactory"/>
<filter class="solr.LowerCaseFilterFactory"/>
<filter class="solr.SnowballPorterFilterFactory" language="English" />
</analyzer>
<analyzer type="query">
<tokenizer class="solr.StandardTokenizerFactory"/>
<filter class="solr.StopFilterFactory" ignoreCase="true" words="stopwords.txt" enablePositionIncrements="true" />
<!-- We will use synonyms only at index time to keep querying fast
<filter class="solr.SynonymFilterFactory" synonyms="synonyms.txt" ignoreCase="true" expand="true"/>
-->
<filter class="solr.LowerCaseFilterFactory"/>
<filter class="solr.SnowballPorterFilterFactory" language="English" />
</analyzer>
</fieldType>
Couple of things to note:
I am using synonyms only at index time, to keep queries fast.
I added KeywordTokenizerFactory, it treats the entire field as a single token, and does not split multi-word synonyms
I added expand="true". If expand is true, a synonym will be expanded to all equivalent synonyms. If it is false, all equivalent synonyms will be reduced to the first in the list.
Query time synonyms are commented out.