Building Fullstack Graph Applications With Neo4j
6 likes1,585 views
This document provides an overview of graph databases and algorithms using Neo4j. It discusses Neo4j's built-in graph algorithms for pathfinding, centrality, community detection, similarity and link prediction. It also covers Neo4j Streams for real-time graph processing and integrations with Kafka. Grandstack and Neo4j-GraphQL are presented as options for building GraphQL APIs on Neo4j.
Building Fullstack Graph Applications With Neo4j
- 1. Fullstack Graph w/ Neo4j William Lyon @lyonwj lyonwj.com bit.ly/seattlegraph Graph Tour Seattle
- 2. William Lyon Software Developer @Neo4j ● Neo4j Labs ■ Integrations ■ GraphQL ■ Data Visualization ■ ... @lyonwj lyonwj.com will@neo4j.com
- 4. Pig E. Bank
- 15. medium.com/neo4j
- 19. ● ● ●
- 22. 1. → 2. 3. 4. 5. Arrows for data model diagrams: www.apcjones.com/arrows/
- 26. ● ● ●
- 45. neo4j.com/labs/
- 48. About APOC • Large standard library of utility functions and procedures • Actively developed - many contributors • "scratch your itch" • Makes Cypher easier to use • Enable some specific use-cases • Plan is it migrate some of the functionality into the Neo4j product
- 49. Browser Guide :play apoc • Explore live examples
- 50. • Relational / Cassandra • MongoDB, Couchbase, ElasticSearch • JSON, XML, CSV, XLS • Cypher, GraphML • … Data integration
- 51. Load from a relational database
- 52. apoc.load.jdbc WITH "jdbc:mysql://localhost:3306/customers?user=root" AS url CALL apoc.load.jdbc(url,"parties") YIELD row MERGE (p:Party {id: row.PartyID}) SET p.name = row.name, p.ssn = row.SSN
- 57. {event} {event}
- 59. {event} {event}
- 61. Graph Algorithms in Neo4j • Parallel Breadth First Search & DFS • Shortest Path • Single-Source Shortest Path • All Pairs Shortest Path • Minimum Spanning Tree • A* Shortest Path • Yen’s K Shortest Path • K-Spanning Tree (MST) • Random Walk • Degree Centrality • Closeness Centrality • CC Variations: Harmonic, Dangalchev, Wasserman & Faust • Betweenness Centrality • Approximate Betweenness Centrality • PageRank • Personalized PageRank • ArticleRank • Eigenvector Centrality • Triangle Count • Clustering Coefficients • Connected Components (Union Find) • Strongly Connected Components • Label Propagation • Louvain Modularity – 1 Step & Multi-Step • Balanced Triad (identification) • Euclidean Distance • Cosine Similarity • Jaccard Similarity • Overlap Similarity • Pearson Similarity Pathfinding & Search Centrality / Importance Community Detection Similarity Updated June 2019 Link Prediction • Adamic Adar • Common Neighbors • Preferential Attachment • Resource Allocations • Same Community • Total Neighbors +35
- 68. ● Fullstack framework for building applications Fullstack GraphQL with GRANDstack grandstack.io @lyonwj
- 69. What is GraphQL? graphql.org An API query language and runtime for building APIs @lyonwj
- 70. “Your Application Data Is A Graph” -- GraphQL @lyonwj
- 72. Movies, Genres, Directors, Actors GraphQL Type Definitions ● Defined using GraphQL Schema Definition Language (SDL) @lyonwj
- 73. Movies, Genres, Directors, Actors Introspection ● Schema can be queried ● Schema becomes API specification / documentation ● Tools like GraphiQL / GraphQL Playground @lyonwj
- 74. Movies, Genres, Directors, Actors GraphQL query @lyonwj
- 77. GraphQL query Movies, Genres, Directors, Actors @lyonwj
- 78. GraphQL query GraphQL response Movies, Genres, Directors, Actors @lyonwj
- 79. 1) GraphQL is an API query language, not a database query language. 2) Limited expressivity (no projections, no aggregations, etc). 3) While GraphQL exposes your application data as a graph, it’s not (just) for graph databases What is GraphQL? @lyonwj
- 80. ● Overfetching ○ Less data over the wire ● Underfetching ○ Single round trip ● GraphQL Specification ● “Graphs All The Way Down” ○ Relationships vs Resources ○ Unify disparate systems (microservices) ● Simplify data fetching ○ Component based data interactions GraphQL Advantages @lyonwj
- 81. ● Some well understood practices from REST don’t apply ○ HTTP status codes ○ Errors ○ Caching ● Exposing arbitrary complexity to client ○ Performance considerations ● n+1 query problem ● Query costing / rate limiting GraphQL Challenges @lyonwj
- 82. Example API & How It’s Built
- 83. Example GraphQL API Neo4j Community Forum Trending Activity Feeds community.neo4j.com @lyonwj
- 85. How To Build A GraphQL Service
- 87. Community Graph http://138.197.15.1:7474/browser/ user: community password: community@lyonwj
- 88. @lyonwj
- 89. Start With A GraphQL Schema graphql.org/learn/schema/#type-language @lyonwj
- 90. Start With A GraphQL Schema Type Definitions @lyonwj
- 91. Start With A GraphQL Schema API Entry Point(s) Query & Mutation Types @lyonwj
- 92. Start With A GraphQL Schema API Entry Point(s) Query & Mutation Types graphql.org/learn/schema/#type-language @lyonwj
- 93. ● Functions that define how to “resolve” data for GraphQL request GraphQL Resolvers graphql.org/learn/execution/#root-fields-resolvers@lyonwj
- 102. 1) Schema Duplication 2) Mapping / translation layer from graph ←→ (???) 3) Boilerplate code 4) n+1 query problem Common Problems With This Approach https://blog.grandstack.io/five-common-graphql-problems-and-how-neo4j-graphql-aims-to-solve-them-e9a8999c8d43 @lyonwj
- 103. GraphQL “Engines”
- 104. ● Tools for auto-generating GraphQL schema, generating database queries from GraphQL requests GraphQL “Engines” Overview prisma.io aws.amazon.com/appsync hasura.io graphile.org grandstack.io Neo4j-GraphQL @lyonwj
- 105. Neo4j-GraphQL
- 106. ● GraphQL First Development ○ GraphQL schema drives the database data model ● Generate Cypher from GraphQL ○ Single query / single round trip to database ● Generate GraphQL CRUD API from type definitions ● Auto-generated resolvers (no boilerplate!) ● Extend GraphQL functionality with Cypher ○ @cypher schema directive Goals for Neo4j-GraphQL Integration @lyonwj
- 108. Auto-generate GraphQL CRUD API @lyonwj Query & Mutation types ○ Entry point for each type Ordering Pagination Complex Filter DateTime types https://blog.grandstack.io/complex-graphql-filtering-with-neo4j-graphql-js-aef19ad06c3e
- 109. Generate Cypher From GraphQL @lyonwj
- 110. Extend GraphQL w/ Cypher grandstack.io/docs/neo4j-graphql-js.html#cypher-directive @cypher GraphQL schema directive @lyonwj
- 115. ● Declarative database integrations for GraphQL ● GraphQL type definitions define database model ● Provision CRUD GraphQL API ○ Auto-generated GraphQL API ○ Schema enrichments ● Generate database queries ○ auto-generated resolvers ○ reduce boilerplate GraphQL ”Engines” @lyonwj
- 116. How Do GraphQL Engines Generate Database Queries From GraphQL Requests? resolveInfo resolver argument ● GraphQL query AST ● GraphQL schema ● Selection set ● Variables ● ... @lyonwj
- 118. GraphQL Deployment “Traditional” Approaches Serverless Options ● PaaS ○ Heroku, AWS Elastic Beanstalk ● Docker container ● VPS ● ... ● AWS Lambda, Google Cloud Functions, Azure ● Serverless Framework ○ serverless.com ● Zeit Now ○ zeit.co/now ● Netlify Functions ○ netlify.com @lyonwj
- 119. neo4j-graphql-js apollo-server Neo4j Cloud AWS Lambda GCP Functions Static CDN GraphQL Deployment @lyonwj
- 121. grandstack.io
- 123. GraphQL Clients Apollo Client Alternatives ● Most popular ● Frontend integrations for: ○ React, Angular, vue.js, scala.js,iOS, Android, etc ● Relay ● urql ● graphql-request ● GraphiQL ● GraphQL Playground ● fetch ● <any http client> @lyonwj
- 124. Apollo Client (for React) @lyonwj
- 125. Apollo Client (for React) @lyonwj
- 126. Apollo Client (for React) @lyonwj
- 127. Query GraphQL endpoint Render our table @lyonwj
- 129. Handle GraphQL response and render table @lyonwj
- 130. @lyonwj
- 133. ● ●
- 135. ● ● ● ● ○ ○ ○
- 136. ● ●
- 137. Fraud Flagger Party Interaction Stream Fraud Suspect Stream Fraud Confirmation Stream neo4j-streams
- 138. Investigation Account Action Improved Insight
- 140. 1. a. 2. 3.
- 141. Graph Algorithm Categories in Neo4j neo4j.com/ graph-algorithms- book/ Pathfinding & Search Centrality / Importance Community Detection Link Prediction Finds optimal paths or evaluates route availability and quality Determines the importance of distinct nodes in the network Detects group clustering or partition options Evaluates how alike nodes are Estimates the likelihood of nodes forming a future relationship Similarity
- 142. Graph Algorithms in Neo4j • Parallel Breadth First Search & DFS • Shortest Path • Single-Source Shortest Path • All Pairs Shortest Path • Minimum Spanning Tree • A* Shortest Path • Yen’s K Shortest Path • K-Spanning Tree (MST) • Random Walk • Degree Centrality • Closeness Centrality • CC Variations: Harmonic, Dangalchev, Wasserman & Faust • Betweenness Centrality • Approximate Betweenness Centrality • PageRank • Personalized PageRank • ArticleRank • Eigenvector Centrality • Triangle Count • Clustering Coefficients • Connected Components (Union Find) • Strongly Connected Components • Label Propagation • Louvain Modularity – 1 Step & Multi-Step • Balanced Triad (identification) • Euclidean Distance • Cosine Similarity • Jaccard Similarity • Overlap Similarity • Pearson Similarity Pathfinding & Search Centrality / Importance Community Detection Similarity Updated June 2019 Link Prediction • Adamic Adar • Common Neighbors • Preferential Attachment • Resource Allocations • Same Community • Total Neighbors +35
- 145. ● ● ●
- 146. ● ● ● ● ●
- 147. ● ○ https://github.com/neo4j-contrib/neo4j-streams ○ https://www.confluent.io/hub/neo4j/kafka-connect-neo4j ● https://grandstack.io/ ● https://www.freecodecamp.org/news/how-to-leverage-neo4j-streams-and-build-a-just-in-time-data-warehouse-64adf290f093/ ● https://neo4j.com/docs/graph-algorithms/current/