I am struggling with a little problem where I have to display relevant information about the resultset returned from SolR but can't figure out how to calculate it without iterating the results (bad).
Basically I am storing my documents with a state field and while the search is supposed to return all documents, the UI has to show "Found 15 entities, 5 are in state A, 3 in state B and 8 in C".
At the moment I am using a rather brittle approach of running the query 3 times with additional scoping by type, but I'd rather get that information from the one query I am displaying. (There have been some edge cases where the numbers don't add up and since SolR can return facets I guess there has to be a way to use that functionality in this case)
I am using SolR 3.5 from Rails with the sunspot gem
As you mention yourself, you can use facets for this by setting
facet=true&facet.field=state
I'm not familiar with the sunspot gem, but by looking at the documentation you can use
facets like this(Assuming Entity is your searchable):
Entity.search do:
facet :state
end
This should return the states of all entities returned by your query with the number of entities in this state. The Sunspot documentation tells me you can read these facets in the following way:
search.facet(:state).rows.each do |facet|
puts "State #{facet.value} has #{facet.count} entities"
end
Essentially there are three main sets of functions you can use to garner stats from solr.
The first is faceting:
http://wiki.apache.org/solr/SimpleFacetParameters
There is also grouping (field collapsing):
https://wiki.apache.org/solr/FieldCollapsing
And the stats package:
https://cwiki.apache.org/confluence/display/solr/The+Stats+Component
Although the stats, facet and group may be replaced by the analytic package known as olap which is aimed to be in solr V 5.0.0:
https://issues.apache.org/jira/browse/SOLR-5302
Good luck.
Related
I'm facing a issue wherein I have huge amount of data in Solr and as a result, searching for a multi token query is generating a big recall set. For ex - if i search for "apple watch series 4 42mm", i get back 4 million results. My parser is edismax, minimum match setting is 2 as of now, and am using WhiteSpace Tokenizer with a bunch of filters. The goal here is to reduce this recall set to display more relevant results.
Things that I explored are -
MinimumMatch - Am trying setting mm to 2>2 4>3 to see how it results. Also tried finding out if i could apply mm on individual fields and found out that it used to be possible with local params in Solr but has been discontinued since Solr 7.2. I do not want to get into writing a custom parser or tweaking Solr's code since that could lead to other problems. Nor do i want to change the default parser to Lucene. Is there any other way that i could apply mm separately to category_name, product_name, product_description, brand_name, etc?
Query slop - Am not using qs as of now, tried a few examples converting my query into phrase query and applying qs. It does reduce recall but i have a problem there. Suppose i have a product which has "apple" in brand_name and "watch series 4 42mm" in product name, that is a relevant result but will not be returned because the phrase query has to have all tokens in the field. Is there a way to apply qs to suit my purpose?
ShingleFilterFactory - I'm trying this filter with outputUnigrams true because i do not want the individual terms to not be indexed. But with that, index size would explode and result set won't be that good either. Can i use other levers like mm or something else along with this to make it work? Also, is there a way to make outputUnigrams a query param?
Explored pf2, pf3, ps also but those will be used for boosting. Right now, my aim is filtering the most relevant results.
Can someone please help me with the above? Thanks
I am attempting to implement a search in solr 5.5 which requires faceting on child document fields. I realize that flattening the data structure is the ideal solution for solr search but unfortunately because of business requirements of the search, I am required to maintain a relationship between various fields (hence the child documents).
I am experimenting with using the BlockJoinFacetComponent to facet on child document fields, and I am able to get everything working and get the counts I expect using the basic example, no problems there. The issue I am facing is that the BlockJoinFacetComponent requires a ToParentQuery, and I can't figure out how to combine this with my original search query and still get facet results.
To explain further:
I am basically following this example: http://www.slideshare.net/lucidworks/faceting-with-lucene-block-join-query-oleg-savrasov
In the example, the user originally searches for "dress", and then is shown facets to filter down by size, color. Size and Color are child fields, and the BlockJoinFacetComponent is used in the example to facet by size and color and retrieve the expected counts.
In the example, the query used to retrieve said facets (slide 22) is:
q= {!parent which="scope:product"} COLOR: Blue
child.facet.field = SIZE
Which works fine. What I am not understanding is in this example we have now lost the original search for "dress". So my question is basically how can I combine my original search (dress) with the ToParentQuery? I have tried everything I can think of to combine the queries, but I always end up getting the same exception:
"Block join faceting is allowed with ToParentBlockJoinQuery only".
I have even downloaded the solr source code and hooked up a remote debugger where this error is being thrown to try and debug this, but I still can't figure it out. No matter what I do it seems like unless the ToParentBlockJoinQuery is the only thing in the query, the BlockJoinFacetComponent will reject it. Which seems odd considering to use the component you've now lost what the user originally searched for.
After further debugging, the issue stems from the fact that BlockJoinFacetComponent seems to not be able to separate the ToParentBlockJoinQuery portion of a query if you are using a query parser other than the standard parser (I was using edismax).
For example, with the standard query parser, this works:
"original query" + _query_:"{!parent which="scope:product"} COLOR: Blue"
child.facet.field = SIZE
If you run this same query with a dismax or edismax query parser, you receive the error:
"Block join faceting is allowed with ToParentBlockJoinQuery only".
Since I am dependent on the edismax query parser, this was a show stopper for me. However, I was able to achieve the results I desired instead by using the JSON Facet API: http://yonik.com/solr-nested-objects/#faceting
I would like to implement relevance feedback in Solr. Solr already has a More Like This feature: Given a single document, return a set of similar documents ranked by similarity to the single input document. Is it possible to configure Solr's More Like This feature to behave like More Like Those? In other words: Given a set of documents, return a list of documents similar to the input set (ranked by similarity).
According to the answer to this question turning Solr's More Like This into More Like Those can be done in the following way:
Take the url of the result set of the query returning the specified documents. For example, the url http://solrServer:8983/solr/select?q=id:1%20id:2%20id:3 returns the response to the query id:1 id:2 id:3 which is practically the concatenation of documents 1, 2, 3.
Put the above url (concatenation of the specified documents) in the url.stream GET parameter of the More Like This handler: http://solrServer:8983/solr/mlt?mlt.fl=text&mlt.mintf=0&stream.url=http://solrServer:8983/solr/select%3Fq=id:1%20id:2%20id:3. Now the More Like This handler treats the concatenation of documents 1, 2 and 3 as a single input document and returns a ranked set of documents similar to the concatenation.
This is a pretty bad implementation: Treating the set of input documents like one big document discriminates against short documents because short documents occupy a small portion of the entire big document.
Solr's More Like This feature is implemented by a variation of The Rocchio Algorithm: It takes the top 20 terms of the (single) input document (the terms with the highest TF-IDF values) and uses those terms as the modified query, boosted according to their TF-IDF. I am looking for a way to configure Solr's More Like This feature to take multiple documents as its input, extract the top n terms from each input document and query the index with those terms boosted according to their TF-IDF.
Is it possible to configure More Like This to behave that way? If not, what is the best way to implement relevance feedback in Solr?
Unfortunately, it is not possible to configure the MLT handler that way.
One way to do it would be to implement a custom SearchComponent and register it to a (dedicated) SearchHadler.
I've already done something similar and it is quite easy if you look a the original implementation of MLT component.
The most difficult part is the synchronization of the results from different shard servers, but it can be skipped if you do not use shards.
I would also strongly recommend to use your own parameters in your implementation to prevent collisions with other components.
Imagine an index like the following:
id partno name description
1 1000.001 Apple iPod iPod by Apple
2 1000.123 Apple iPhone The iPhone
When the user searches for "Apple" both documents would be returned. Now I'd like to give the user the possibility to narrow down the results by limiting the search to one or more fields that have documents containing the term "Apple" within those fields.
So, ideally, the user would see something like this in the filter section of the ui after his first query:
Filter by field
name (2)
description (1)
When the user applies the filter for field "description", only documents which contain the term "Apple" within the field "description" would be returned. So the result set of that second request would be the iPod document only. For that I'd use a query like ?q=Apple&qf=description (I'm using the Extended DisMax Query Parser)
How can I accomplish that with Solr?
I already experimented with faceting, grouping and highlighting components, but did not really come to a decent solution to this.
[Update]
Just to make that clear again: The main problem here is to get the information needed for displaying the "Filter by field" section. This includes the names of the fields and the hits per field. Sending a second request with one of those filters applied already works.
Solr just plain Doesn't Do This. If you absolutely need it, I'd try it the multiple requests solution and benchmark it -- solr tends to be a lot faster than what people put in front of it, so an couple few requests might not be that big of a deal.
you could achieve this with two different search requests/queries:
name:apple -> 2 hits
description:apple -> 1 hit
EDIT:
You also could implement your own SearchComponent that executes multiple queries in the background and put it in the SearchHandler processing chain so you only will need a single query in the frontend.
if you want the term to be searched over the same fields every time, you have 2 options not breaking the "single query" requirement:
1) copyField: you group at index time all the fields that should match togheter. With just one copyfield your problem doesn't exist, if you need more than one, you're at the same spot.
2) you could filter the query each time dynamically adding the "fq" parameter at the end
http://<your_url_and_stuff>/?q=Apple&fq=name:Apple ...
this works if you'll be searching always on the same two fields (or you can setup them before querying) otherwise you'll always need at least a second query
Since i said "you have 2 options" but you actually have 3 (and i rushed my answer), here's the third:
3) the dismax plugin described by them like this:
The DisMaxQParserPlugin is designed to process simple user entered phrases
(without heavy syntax) and search for the individual words across several fields
using different weighting (boosts) based on the significance of each field.
so, if you can use it, you may want to give it a look and start from the qf parameters (that is what the option number 2 wanted to be about, but i changed it in favor of fq... don't ask me why...)
SolrFaceting should solve your problem.
Have a look at the Examples.
This can be achieved with Solr faceting, but it's not neat. For example, I can issue this query:
/select?q=*:*&rows=0&facet=true&facet.query=title:donkey&facet.query=text:donkey&wt=json
to find the number of documents containing donkey in the title and text fields. I may get this response:
{
"responseHeader":{"status":0,"QTime":1,"params":{"facet":"true","facet.query":["title:donkey","text:donkey"],"q":"*:*","wt":"json","rows":"0"}},
"response":{"numFound":3365840,"start":0,"docs":[]},
"facet_counts":{
"facet_queries":{
"title:donkey":127,
"text:donkey":4108
},
"facet_fields":{},
"facet_dates":{},
"facet_ranges":{}
}
}
Since you also want the documents back for the field-disjunctive query, something like the following works:
/select?q=donkey&defType=edismax&qf=text+titlle&rows=10&facet=true&facet.query=title:donkey&facet.query=text:donkey&wt=json
I'm starting to think that what I want to do is not possible but thought I would give this a try.
I'm running Solr 3.5.
I currently have two types of search:
A basic spatial query which returns the calulated distance between two points in the score field.
Sample Query from my Solr logs:
?fl=*,score&sort=score+asc&start=0&q={!func}geodist()&sfield=coordinates&pt=59.2363514,18.092783&version=2
A dismax query which allows free text queries on a number of fields.
Sample Query from Solr log:
mm=1&d=100.0&sfield=coordinates&qf=field1^5.0+fields2^3.0&defType=edismax&version=2&fl=*,score&start=1&q=monkeyhopper&pt=59.2363514,18.0927830000&fq={!geofilt}}
I want to replace my first query with the dismax query but I really need to get the calculated distance in the response. Yes, I can calulate the distance programatically but I would prefer not having to do this as Solr has done it for me already.
I still want to be able to sort my dismax query "by relevance", distance or any other field so the score given by my boosts could be interesting for sorting but I don't need it to be returned.
If I understood correctly you want to have the result of a function in your Solr response. The SOLR-2444 issue is what you're looking for I guess: it allows to include in the fl parameter pseudo-fields, functions etc. The only problem is that it's been committed only on trunk, so it isn't available on the current Solr release, neither will be in the coming 3.6 release. You have to wait for the 4 release but I don't think it will take a lot of time. Maybe you can already start playing around with a snapshot of the last successful Jenkins build.
Pseudo-fields are now available in Solr 4+ which allow you to do just this.
http://localhost:8983/solr/collection1/browse?q=*:*&rows=1000&wt=xml&pt=37.763649,-122.24313&sfield=store&fl=dist:geodist()
For instance, this request allows me to return a field "dist" which contains the distance of each entry to the stated point.