Merge pull request #2559 from WheresMyStapler/text-predicate-negations

Add negations to all text predicates

docs/changelog.md

@@ -277,6 +277,32 @@ Following, classes are removed and have to be replaced by tinkerpop equivalent:
 | `org.janusgraph.channelizers.JanusGraphNioChannelizer` | `org.apache.tinkerpop.gremlin.server.channel.NioChannelizer` |
 | `org.janusgraph.channelizers.JanusGraphWsAndHttpChannelizer` | `org.apache.tinkerpop.gremlin.server.channel.WsAndHttpChannelizer` |
 
+##### Breaking change Lucene and Solr fuzzy predicates
+
+The text predicates `text.textFuzzy` and `text.textContainsFuzzy` have been updated in both the Lucene and Solr indexing
+backends to align with JanusGraph and Elastic. These predicates now inspect the query length to determine the Levenshtein
+distance, where previously they used the backend's default max distance of 2:
+
+- 0 for strings of one or two characters (exact match)
+- 1 for strings of three, four or five characters
+- 2 for strings of more than five characters
+
+**Change Matrix:**
+
+| text | query | previous result | new result |
+| --- | --- | --- | --- |
+| ah | ah | true | true |
+| ah | ai | true | **false** |
+| hop | hop | true | true |
+| hop | hap | true | true |
+| hop | hoop | true | true |
+| hop | hooop | true | **false** |
+| surprises | surprises | true | true |
+| surprises | surprizes | true | true |
+| surprises | surpprises | true | true |
+| surprises | surpprisess | false | false |
+
+
 ### Version 0.5.3 (Release Date: December 24, 2020)
 
 === "Maven"

docs/index-backend/text-search.md

@@ -68,6 +68,15 @@ g.V().has('booksummary', textContainsRegex('.*corn.*'))
 g.V().has('booksummary', textContainsFuzzy('unicorn'))
 ```
 
+The Elasticsearch backend extends this functionality and includes support for negations
+of the above predicates, as well as phrase matching:
+
+- `textNotContains`: is true if no words inside the text string match the query string
+- `textNotContainsPrefix`: is true if no words inside the text string begin with the query string
+- `textNotContainsRegex`: is true if no words inside the text string match the given regular expression
+- `textNotContainsFuzzy`:  is true if no words inside the text string are similar to the query string (based on Levenshtein edit distance)
+- `textNotContainsPhrase`:  is true if the text string does not contain the sequence of words in the query string
+
 String search predicates (see below) may be used in queries, but those
 require filtering in memory which can be very costly.
 
@@ -111,6 +120,13 @@ g.V().has('bookname', textRegex('.*corn.*'))
 g.V().has('bookname', textFuzzy('unicorn'))
 ```
 
+The Elasticsearch backend extends this functionality and includes support for negations
+of the above text predicates:
+
+- `textNotPrefix`: if the string value does not start with the given query string
+- `textNotRegex`: if the string value does not match the given regular expression in its entirety
+- `textNotFuzzy`: if the string value is not similar to the given query string (based on Levenshtein edit distance)
+
 Full-text search predicates may be used in queries, but those require
 filtering in memory which can be very costly.
 

docs/interactions/search-predicates.md

@@ -27,13 +27,22 @@ The `Text` enum specifies the [Text Search](../index-backend/text-search.md) use
 
 * Text search predicates which match against the individual words inside a text string after it has been tokenized. These predicates are not case sensitive.
     - `textContains`: is true if (at least) one word inside the text string matches the query string
+    - `textNotContains`: is true if no words inside the text string match the query string
     - `textContainsPrefix`: is true if (at least) one word inside the text string begins with the query string
+    - `textNotContainsPrefix`: is true if no words inside the text string begin with the query string
     - `textContainsRegex`: is true if (at least) one word inside the text string matches the given regular expression
-    - `textContainsFuzzy`: is true if (at least) one word inside the text string is similar to the query String (based on Levenshtein edit distance)
+    - `textNotContainsRegex`: is true if no words inside the text string match the given regular expression
+    - `textContainsFuzzy`: is true if (at least) one word inside the text string is similar to the query string (based on Levenshtein edit distance)
+    - `textNotContainsFuzzy`:  is true if no words inside the text string are similar to the query string (based on Levenshtein edit distance)
+    - `textContainsPhrase`: is true if the text string contains the exact sequence of words in the query string
+    - `textNotContainsPhrase`:  is true if the text string does not contain the sequence of words in the query string
 * String search predicates which match against the entire string value
     - `textPrefix`: if the string value starts with the given query string
+    - `textNotPrefix`: if the string value does not start with the given query string
     - `textRegex`: if the string value matches the given regular expression in its entirety
+    - `textNotRegex`: if the string value does not match the given regular expression in its entirety
     - `textFuzzy`: if the string value is similar to the given query string (based on Levenshtein edit distance)
+    - `textNotFuzzy`: if the string value is not similar to the given query string (based on Levenshtein edit distance)
 
 See [Text Search](../index-backend/text-search.md) for more information about full-text and string search.