Discovering graph structures

Graph is a mathematical construct used to model the relationships between key/value pairs, comprises a set of vertices (nodes) and an arbitrary number of edges (lines) which connect them. There are many ways to implement graphs with PHP, from typical SQL solutions to dedidacted, powerful tools like Apache Giraph or Titan. In this talk I will introduce Neo4j, highly scalable and robust native graph database, used in mission-critical apps by thousands of leading startups, enterprises and governments around the world. We will also talk about managing nodes and traversing methods with PHP.

Transcript

1. DISCOVERING GRAPH STRUCTURES
   Mariusz Gil
   

   View Slide

2. ABOUT ME
   

   View Slide

3. DEVELOPER / TECH-LEAD / CONSULTANT
   

   View Slide

4. View Slide

5. INSPIRATION
   IS BETTER THAN SOLUTION
   

   View Slide

6. DATA
   

   View Slide

7. CRUCIAL
   VERY THING IN YOUR APP
   

   View Slide

8. CRUCIAL
   MOST THING IN YOUR APP
   STORING PROCESSING
   

   View Slide

9. 90+
   

   View Slide

10. WHY?
    

    View Slide

11. 90+ HYPOTHESIS
    

    View Slide

12. dev
    developers
    struct
    set
    list
    tree
    graph
    …
    SQL(struct)
    

    View Slide

13. WE ALL SPEAK SQL*
    * INCLUDING 3NF
    THIS IS OUR WAY TO WORK WITH DATA
    

    View Slide

14. 7MB QUERY
    

    View Slide

15. STORING PROCESSING
    

    View Slide

16. 90+ HYPOTHESIS EXTENDED
    

    View Slide

17. dev
    developers
    struct
    set
    list
    tree
    graph
    …
    SQL(JSON(struct))
    

    View Slide

18. View Slide

19. View Slide

20. View Slide

21. View Slide

22. GRAPHS
    

    View Slide

23. POWERFUL SINCE 1736
    

    View Slide

24. C
    A
    GRAPH = (VERTICES, EDGES)
    B
    

    View Slide

25. C
    A
    SPARSE
    GRAPH
    B
    D
    C
    A
    COMPLETE
    GRAPH
    B
    D
    C
    A
    DENSE
    GRAPH
    B
    D
    ADJACENCY LIST
    ADJACENCY MATRIX
    

    View Slide

26. DIGRAPH = (VERTICES, DIRECTED EDGES)
    C
    A
    B
    

    View Slide

27. type: KNOWS
    VERTICES AND EDGES PROPERTIES
    B
    type: NOTHING
    A
    type: PERSON
    name: John Snow
    

    View Slide

28. GRAPHS IN REAL LIFE
    

    View Slide

29. View Slide

30. Open
    In progress
    Resolved
    Reopened Closed
    

    View Slide

31. task_id task_status
    1 1
    2 3
    3 2
    4 4
    Open
    In progress
    Resolved
    Reopened Closed
    

    View Slide

32. use Finite\StatefulInterface;
    class Document implements StatefulInterface
    {
    private $state;
    public function setFiniteState($state)
    {
    $this->state = $state;
    }
    public function getFiniteState()
    {
    return $this->state;
    }
    }
    

    View Slide

33. use Finite\StateMachine\StateMachine;
    use Finite\State\State;
    use Finite\State\StateInterface;
    // $document = retrieve your stateful object
    $sm = new StateMachine();
    // Define states
    $sm->addState(
    new State('s1', StateInterface::TYPE_INITIAL)
    );
    $sm->addState('s2');
    $sm->addState('s3');
    $sm->addState(
    new State('s4', StateInterface::TYPE_FINAL)
    );
    

    View Slide

34. s4
    s3
    s2
    // Define transitions
    $sm->addTransition('t12', 's1', 's2');
    $sm->addTransition('t23', 's2', 's3');
    $sm->addTransition('t34', 's3', 's4');
    $sm->addTransition('t42', 's4', 's2');
    // Initialize
    $sm->setObject($document);
    $sm->initialize();
    // Retrieve current state
    $sm->getCurrentState();
    // Can we process a transition?
    $sm->can('t34');
    s1
    

    View Slide

35. HUNDREDS TRANSITIONS
    TENS OF MODELS
    ONE ACTION
    DRY
    HUNDREDS OF THOUSANDS OBJECTS IN DB
    

    View Slide

36. /**
    * @Route("/change-state")
    */
    public function changeStateAction() {
    // Get model
    // ...
    // Check user permissions
    // ...
    // Check transition and notify observers
    if ($model->isTransitionAllowed($targetState) {
    $model->setState($targetState);
    // Save change
    // ...
    }
    // Finish
    // ...
    }
    

    View Slide

37. E
    C
    G
    F
    B
    A
    D
    H
    I
    J
    K
    L
    M
    

    View Slide

38. E
    C
    G
    F
    B
    A
    D
    H
    I
    J
    1:30
    1:30
    1:00
    1:45
    0:45
    0:45
    3:00
    1:15
    1:00
    0:40
    0:20
    0:30
    0:30
    K
    L
    M
    

    View Slide

39. @tom
    @bob
    @john
    #php
    #js
    #web
    Tweet X
    search
    tweet
    retweet
    favourite
    contains
    search
    search
    search
    follow
    follow
    search
    contains
    search
    

    View Slide

40. require_once 'vendor/autoload.php';
    use \Fhaculty\Graph\Graph as Graph;
    $graph = new Graph();
    // Create some cities
    $rome = $graph->createVertex('Rome');
    $madrid = $graph->createVertex('Madrid');
    $cologne = $graph->createVertex('Cologne');
    //Bbuild some roads
    $cologne->createEdgeTo($madrid);
    $madrid->createEdgeTo($rome);
    // create loop
    $rome->createEdgeTo($rome);
    // Process
    foreach ($rome->getVerticesEdgeFrom() as $vertex) {
    echo $vertex->getId().' leads to rome’ . PHP_EOL;
    }
    

    View Slide

41. require_once 'vendor/autoload.php';
    $graph = new Fhaculty\Graph\Graph();
    $blue = $graph->createVertex('blue');
    $blue->setAttribute('graphviz.color', 'blue');
    $red = $graph->createVertex('red');
    $red->setAttribute('graphviz.color', 'red');
    $edge = $blue->createEdgeTo($red);
    $edge->setAttribute('graphviz.color', 'grey');
    $graphviz = new Graphp\GraphViz\GraphViz();
    $graphviz->display($graph);
    

    View Slide

42. SEARCH
    SHORTEST PATH
    SPANNING TREES
    TRAVELING SALESMAN PROBLEM
    MAXIMUM FLOW
    MINIMUM COST FLOW
    MAXIMUM MATCHING
    ALGORITHMS
    

    View Slide

43. TYPICAL SQL-WAY
    C
    A
    B
    mysql> SELECT * FROM `nodes`;
    +----+------+
    | id | name |
    +----+------+
    | 1 | A |
    | 2 | B |
    | 3 | C |
    +----+------+
    3 rows in set (0,00 sec)
    mysql> SELECT * FROM `edges`;
    +--------+--------+
    | s_node | e_node |
    +--------+--------+
    | 1 | 2 |
    | 2 | 3 |
    | 3 | 1 |
    +--------+--------+
    3 rows in set (0,00 sec)
    CREATE TABLE `nodes` (
    `id` int(11) NOT NULL auto_increment,
    `name` varchar(64) NOT NULL,
    PRIMARY KEY (`id`)
    ) ENGINE=InnoDB;
    CREATE TABLE `edges` (
    `s_node` int(11) NOT NULL,
    `e_node` int(11) NOT NULL,
    PRIMARY KEY (`s_node`,`e_node`),
    ) ENGINE=InnoDB;
    

    View Slide

44. TYPICAL SQL-WAY
    C
    A
    B
    mysql> SELECT * FROM `nodes`;
    +----+------+
    | id | name |
    +----+------+
    | 1 | A |
    | 2 | B |
    | 3 | C |
    +----+------+
    3 rows in set (0,00 sec)
    mysql> SELECT * FROM `edges`;
    +--------+--------+-------+
    | s_node | e_node | float |
    +--------+--------+-------+
    | 1 | 2 | 6.50 |
    | 2 | 3 | 4.75 |
    | 3 | 1 | 10.00 |
    +--------+--------+-------+
    3 rows in set (0,00 sec)
    CREATE TABLE `nodes` (
    `id` int(11) NOT NULL auto_increment,
    `name` varchar(64) NOT NULL,
    PRIMARY KEY (`id`)
    ) ENGINE=InnoDB;
    CREATE TABLE `edges` (
    `s_node` int(11) NOT NULL,
    `e_node` int(11) NOT NULL,
    `weight` float NOT NULL,
    PRIMARY KEY (`s_node`,`e_node`),
    ) ENGINE=InnoDB;
    10.00
    6.50 4.75
    

    View Slide

45. View Slide

46. LARGE
    MEDIUM
    SMALL
    

    View Slide

47. GRAPH ORIENTED DATABASE
    

    View Slide

48. INSTALL
    

    View Slide

49. SINGLE NODE SETUP
    wget -O - http://debian.neo4j.org/neotechnology.gpg.key | apt-key add -
    echo 'deb http://debian.neo4j.org/repo stable/' > /etc/apt/sources.list.d/neo4j.list
    apt-get update
    apt-get install neo4j
    service neo4j-service status
    LINUX EDITION
    

    View Slide

50. View Slide

51. RUN
    

    View Slide

52. View Slide

53. QUERY
    

    View Slide

54. View Slide

55. LIVE DEMO
    

    View Slide

56. require_once 'vendor/autoload.php';
    use Neoxygen\NeoClient\ClientBuilder;
    $client = ClientBuilder::create()
    ->addConnection('default')
    ->setAutoFormatResponse(true)
    ->build();
    $query = 'MATCH (n)-[r]->(m) RETURN *';
    $nodes = $client->sendCypherQuery($query)
    ->getResult()
    ->getNodes();
    foreach ($nodes as $node) {
    echo sprintf(
    "%s %s:" . PHP_EOL,
    $node->getLabel(),
    $node->getProperty("name")
    );
    }
    

    View Slide

57. View Slide

58. http://graphdatabases.com
    

    View Slide

59. ALTERNATIVES
    

    View Slide

60. View Slide

61. View Slide

62. View Slide

63. TOOLS WILL CHANGE
    IDEAS ARE IMMORTAL
    

    View Slide

64. Dijkstra's algorithm
    is an algorithm for finding
    ______________ in graph
    

    View Slide

65. @mariuszgil
    THANKS!
    https://joind.in/14886
    

    View Slide