WWC Connect BLR

Transcript

1. - Sarah Masud, Red Hat Gremlin The universal graph language
2. Nodes: M-Member c - City C - Country XYZ- Organization Edges: D - Director F - Founder B - BoD W- Works at 50K+ -Members 60 -Cities 20 -Countries
3. Agenda 1. What is a graph database? 2. What about our dear RDBMS 3. What the hell is gremlin? 4. Features and use cases of gremlin 5. Demo Women Who Code
4. What is Graph DB? ❖A graph is a collection of vertices and edge. G=(V,E) ❖A graph DB is a system that helps us store, retrieve and consume data in a graph structured format ❖USP- Relationship Women Who Code Vertex 2 [properties] Vertex 1 [properties] Edge [properties]
5. Examples Women Who Code
6. WWC Graph - A Dummy ❖ Entities ➢ People ➢ Cities, Countries ➢ Organizations ❖ Relations ➢ Person -- founder---WWC ➢ Person -- BOD---WWC ➢ Person---Director---City ➢ Person--Member---City ➢ Person--Knows--Person Women Who Code
7. Why not our dear RDBMS :/ Women Who Code Name Founder A Yes B Yes C D E F G 1. In order to find the founders, must have a column with most rows empty 2. In order to map names to cities, perform costly join operations 3. RDBMS is useful for transactional data 4. How to map real time relations between person A and person B?
8. So… What the hell is Gremlin?? Framework: TinkerPop DSL: Gremlin Wrapper: Gremlin-python Driver: SQL-Gremlin
9. Features of Gremlin • Provides backend transparency BECAUSE • Does NOT provide schema management queries • Based on functional programming • Returns the result as an iterator • Default- Edges are directional • Support all primitive Java data types out of the box • +Many more!
10. The Universality of Gremlin
11. Use Cases for TinkerPop-Gremlin 1. Experiment with different graph stores without affecting your query logic 2. Switch context between OLAP and OLTP 3. Support from open source + enterprise
12. Pros and Cons with Gremlin Pros ❖ Open source! ❖ Vendor agnostic ❖ Wrapper and driver support for many languages ❖ Intuitive Cons ❖ Just a DSL not a programming language ❖ No out of box support for schema and indexing ❖ Wrappers are slow
13. Example Query 1 Who are the founders and Board of directors of WWC? Step 1: Go to the WWC org node Step 2 : Find out those people who are connected with the WWC org Step 3: Filter out those people who have an edge with label founder or BOD g.V().hasLabel(‘Org).has(‘name’,’WWC’).inE(‘founder_of’,bod_at’).outV().values(‘name’)
14. Nodes: M-Member c - City C - Country XYZ- Organization Edges: D - Director F - Founder B - BoD W- Works at
15. Example Query 2 WWC Directors who are less than 30 years old Step 1: Find all the people who are WWC directors Step2: Filter them by property age<30 g.V().hasLabel(‘Person’).outE(‘director_of’).inV().has(‘age’,lt(30)).values(‘name’)
16. Nodes: M-Member c - City C - Country XYZ- Organization Edges: D - Director F - Founder B - BoD W- Works at
17. How can Gremlin help WWC? 1. Suggest friends on social media to members of same chapter 2. page Rank the most influential people in WWC network 3. Track growth of chapters 4. Personalize job recommendation 5. Mentor-Mentee opportunity Awesome WWC
18. References 1. TinkerPop 2. Tinkerpop -Gremlin Getting started 3. Gremlin-Titan-Dynamo Setup 4. Gremlin -Python 5. Demo Source Code
19. Q/A THANK YOU