Graph databases for fun and profit (Geekcamp.sg)

Transcript

1. Graph databases for fun and profit Vivek Prahlad
2. About me
5. Contributor to Neo4j.rb
6. Multitenancy, Versioning, Performance
7. Author of Frankenstein
8. A graph primer What can you use graph databases for? Examples (Neo4j.rb)
9. A graph primer What can you use graph databases for? Examples (Neo4j.rb)
10. What is a graph?
11. A graph is an abstract representation of a set of objects where some pairs of objects are connected by links
12. Graphs are all around us
13. Object graphs
14. Product Order Customer places 1: n chooses 1: n
17. http://tambrahmrage.tumblr.com/post/8594230361/rage-against-the-macchinar
24. Undirected Graph
25. Directed Graph
26. Knows Knows
27. Knows Knows Name: Vivek Name: Paul Name: Andy
28. Knows Knows Name: Vivek Name: Paul Name: Andy Since: April 2012
29. Knows Knows Name: Vivek Name: Paul Name: Andy Since: April 2012 Since: September 2003
30. What on earth is a graph database?
31. A database with an explicit graph structure
32. Each node knows it’s neighbours
33. Typically directed graphs
35. Two main operations
36. Find nodes
37. Either explicitly
38. Node[node_id]
39. Or use an index
40. Car.find(“name: Ferr*”)
41. Traverse, given an origin
43. Graph algorithms
44. Shortest path
45. 14 2 11 15 14 9 6 9 10
46. Graph database options
49. The examples will use Neo4j
50. What is Neo4j?
51. Neo4j is an ACID compliant graph database
52. “Graph databases like Neo4j are an esoteric but powerful member of the NoSQL family”
53. Capacity?
54. 32 Billion nodes 32 Billion relationships 64 Billion properties
55. Threadsafe
56. Single depth transactions
57. Nested transactions: ‘Placebo’ transactions
58. Lucene integration
59. TWO query languages
60. Cypher
61. Gremlin
62. Advanced and enterprise editions for monitoring, replication, hot backup
63. A graph primer What can you use graph databases for? Neo4j.rb
64. Typical applications
68. Recommendation systems
70. Fraud detection
71. Geospatial analysis
73. But
74. Most complex domains
75. Involve an object graph
76. A graph primer What can you use graph databases for? Examples (Neo4j.rb)
77. Modeling graphs is important
79. Partitioning data crucial
80. Neo4j Language bindings
81. Java, C#, Ruby, Node.js
82. Embedded vs. REST server
83. Runs on JRuby
84. Builds on ActiveModel
85. validations, nested attributes, composed_of, all work
86. Built in migration support
87. Multitenancy support
88. Versioning
89. 3 layer API
90. Layer 1: Nodes, Relationships
91. Neo4j::Transaction run do Neo4j::Node.new(:name => “A”) end
92. Layer 2: Node and Relationship Mixins
93. class SportsCar include Neo4j::NodeMixin property :brand end
94. Layer 3: Rails Model support
95. class SportsCar < Neo4j::Rails::Model property :brand end
96. class SportsCar < Neo4j::Rails::Model include Neo4j::Rails::Versioning property :brand end
97. class SportsCar < Neo4j::Rails::Model include Neo4j::Rails::Versioning property :brand, :index =>:fulltext end
98. class User < Neo4j::Rails::Model attr_accessor :password attr_accessible :email, :password, :password_confirmation, :pending_account after_save :encrypt_password email_regex = /\A[\w+\-.]+@[a-z\d\-.]+\.[a-z]+\z/i # add an exact lucene index on the email property property :email, :index => :exact has_one(:avatar).to(Avatar) validates :email, :presence => true,:format => { :with => email_regex } validates :email, :uniqueness => true, :unless => :pending_account? accepts_nested_attributes_for :avatar, :allow_destroy => true end
99. Versioning
105. Multitenancy
108. A few cypher examples
109. START a=node(3) MATCH (a)-[:KNOWS]->(b)-[:KNOWS]->(c) RETURN a,b,c
110. |a | b | c | |Node[3]{name->Vivek}|Node[2] |Node[3]|
111. START d=node(1), e=node(2) MATCH p = shortestPath( d-[*..15]->e ) RETURN p
112. (1)--[KNOWS,2]-->(3)--[KNOWS,0]-->(4)--[KNOWS,3]-->(2)
113. In conclusion
114. Graphs are a compelling way to store interconnected data
115. No O/R mismatch
116. More widely applicable than you’d think
117. Try it out!
118. https://github.com/andreasronge/ neo4j.git
119. Thank You
120. @vivekprahlad vprahlad@thoughtworks.com