JanusGraph Text Predicates¶

JanusGraph provides APIs to query the data using Text Predicates. The types of Text Predicates available for querying depends on the Indexing backend used. While using Elasticsearch as Indexing backend exposes all types of Text Predicates, but using some other backend like Solr won’t expose all the Text Predicates. Hence, be careful with using the Text Predicates, and refer to original JanusGraph docs.

Available Text Predicates¶

The Text Predicates are broadly divided into 2 categories as:

Text search predicates which match against the individual words inside a text string after it has been tokenized.
These predicates are not case sensitive.
String search predicates which match against the entire string value

Note

All the bellow examples are based on Graph of the Gods graph which comes pre-packaged with JanusGraph

TextContains¶

This Predicate is part of Text package. TextContains is used to query for partial matches in values of properties. The query is to find all edges where reason contains breezes

from janusgraph_python.core.attribute.TextPredicate.Text import Text

# Find all edges which contains the word breezes
edges = g.E().has("reason", Text.textContains("breezes")).next()
print(edges)

==> e[55m-6hc-b2t-3bk][8400-lives->4304]

# Find all edges which contains the word breezes and loves
edges = g.E().has("reason", Text.textContains("breezes")).has("reason", Text.textContains("loves")).next()
print(edges)

==> e[55m-6hc-b2t-3bk][8400-lives->4304]

TextContainsFuzzy¶

This Predicate is part of Text package. TextContainsFuzzy is used to query for partial matches with Fuzzy logic in values of properties. We know there is a vertex with Hercules value on property name. We query all vertices with value as Herculies Similarly, we use in Conjugation with multiple other attributes to query edge as in above example.

from janusgraph_python.core.attribute.TextPredicate.Text import Text

hercules = g.V().has("name", Text.textContainsFuzzy("hercauleas")).valueMap(True).next()
print(hercules)

==> {'name': ['hercules'], <T.id: 1>: 4312, <T.label: 3>: 'demigod', 'age': [30]}

edges = g.E().has("reason", Text.textContainsFuzzy("breezs")).has("reason", Text.textContains("loves")).next()
print(edges)

==> e[55m-6hc-b2t-3bk][8400-lives->4304]

edges = g.E().has("reason", Text.textContainsFuzzy("breezs")).has("reason", Text.textContainsFuzzy("luves")).next()
print(edges)

==> e[55m-6hc-b2t-3bk][8400-lives->4304]

TextContainsPrefix¶

This Predicate is part of Text package. TextContainsPrefix is used to query for where the value (if sentence, it is segmented, & any word in segment is considered) contains the string queries in prefix The difference between TextContains and TextContainsPrefix and TextPrefix can be seen in example bellow.

from janusgraph_python.core.attribute.TextPredicate.Text import Text

edges = g.E().has("reason", Text.textContainsFuzzy("breezs")).has("reason", Text.textContains("loves")).
        valueMap().toList()
print(edges)

==> [{'reason': 'loves fresh breezes'}]

edges = g.E().has("reason", Text.textContainsPrefix("loves")).valueMap().toList()
print(edges)

==> [{'reason': 'loves fresh breezes'}, {'reason': 'loves waves'}]

edges = g.E().has("reason", Text.textContainsPrefix("breeze")).valueMap().toList()
print(edges)

==> [{'reason': 'loves fresh breezes'}]

edges = g.E().has("reason", Text.textPrefix("breeze")).valueMap().toList()
print(edges)

==> []

TextContainsRegex¶

This Predicate is part of Text package. TextContainsRegex matches with each individual words in string, according to regex provided.

from janusgraph_python.core.attribute.TextPredicate.Text import Text

edges = g.E().has("reason", Text.textContainsRegex("br[ez]*s")).valueMap().toList()
print(edges)

==> [{'reason': 'loves fresh breezes'}]

TextFuzzy¶

This Predicate is part of Text package. TextFuzzy is used to query for complete matches with Fuzzy logic in values of properties. The difference between TextContainsFuzzy and TextFuzzy is seen in example bellow.

from janusgraph_python.core.attribute.TextPredicate.Text import Text

edges = g.E().has("reason", Text.textContainsFuzzy("breezs")).next()
print(edges)

==> e[55m-6hc-b2t-3bk][8400-lives->4304]

edges = g.E().has("reason", Text.textFuzzy("luves fresh breezs")).next()
print(edges)

==> e[55m-6hc-b2t-3bk][8400-lives->4304]

TextPrefix¶

This Predicate is part of Text package. TextPrefix returns the objects, where the string being queries is at the beginning and at prefix of the property’s value. The difference between TextContainsPrefix and TextPrefix is shows in example bellow.

from janusgraph_python.core.attribute.TextPredicate.Text import Text

edges = g.E().has("reason", Text.textContainsPrefix("breeze")).valueMap().toList()
print(edges)

==> [{'reason': 'loves fresh breezes'}]

edges = g.E().has("reason", Text.textContainsPrefix("loves")).valueMap().toList()
print(edges)

==> [{'reason': 'loves fresh breezes'}, {'reason': 'loves waves'}]

edges = g.E().has("reason", Text.textPrefix("breeze")).valueMap().toList()
print(edges)

==> []

TextRegex¶

This Predicate is part of Text package. TextRegex matches the whole string, according to regex provided.

from janusgraph_python.core.attribute.TextPredicate.Text import Text

edges = g.E().has("reason", Text.textRegex('l[ov]*es\s*w[a-v]*')).valueMap().toList()
print(edges)

==> [{'reason': 'loves waves'}]