#Solr suggester example movie#

All lower order models are also built, so if you specify N=3, you will get trigrams, bigrams and unigrams, all compiled into a single weighted FST for maximum sharing of the text tokens. You specify the order (N) of the model when you create the suggester: larger values of N require more data to train properly but can make more accurate predictions. In this case, FreeTextSuggester will see '3d' and the 'p' prefix for the next token and predict printer, even though 'flashforge 3d printer' was never explicitly added as a suggestion. Yet, '3d printer' was a frequently seen phrase in other contexts (different brands). Rather than precisely matching a previous suggestion, it builds up a simple statistical n-gram language model from all suggestions and looks at the last tokens (plus the prefix of whatever final token the user is typing, if present) to predict the most likely next token.įor example, perhaps the user's query so far is: 'flashforge 3d p', and because flashforge is an uncommon brand of 3D printer, this particular suggestion prefix was never added to the suggester.

#Solr suggester example movie#

Alternatively, if your universe of suggestions is inherently closed, such as the movie and show titles that Netflix's search will suggest, or all product names on an e-commerce site, then a closed set of suggestions is appropriate.įor everyone else, where a high percentage of the incoming queries fall into the never-seen-before long tail, Lucene's newest suggester, FreeTextSuggester, can help! It uses the approach described in this Google blog post. This may not be a problem if your suggestions are past user queries and you have tons and tons of them (e.g., you are Google). Yet, a common limitation to all of these suggesters is that they can only suggest from a finite set of previously built suggestions. It is also the only suggester implementation so far that supports highlighting (this has proven challenging for the other suggesters). You can see this one action at the Lucene/Solr Jira search application (e.g., try 'python') that I recently created to eat our own dog food. Lucene's suggest module offers a number of fun auto-suggest implementations to give a user live search suggestions as they type each character into a search box.įor example, WFSTCompletionLookup compiles all suggestions and their weights into a compact Finite State Transducer, enabling fast prefix lookup for basic suggestions.ĪnalyzingSuggester improves on this by using an Analyzer to normalize both the suggestions and the user's query so that trivial differences in whitespace, casing, stop-words, synonyms, as determined by the analyzer, do not prevent a suggestion from matching.įinally, AnalyzingInfixSuggester goes further by allowing infix matches so that words inside each suggestion (not just the prefix) can trigger a match.