This is a demo repo demonstrating how to use Python to interact with Azure Cosmos DB for Apache Gremlin. I prepared a sample dataset of recipes and corresponding ingredients to illustrate their graph relationship. It contains 6 recipes and 16 ingredients.
Each of the recipes & ingredients represents the vertex (aka node). Vertexes are connected via edges representing the relationships (requires). And each vertex and edge has its corresponding property for more descriptive details.
I used "id" in the vertex property to store the recipe's name and ingredient for easier illustration. In the edge property, I used "weight" to illustrate whether or not the ingredient must be required. And used the "quantity" property to illustrate the amount required, e.g., ".property('quantity',3)" of 'Egg' represents 3 eggs are required; ".property('quantity',2)" of 'Cold Rice' represents the 2 cups of cold rice are required; "property('quantity',1)" of 'Soy Sauce' or 'Salt' or 'Sugar' represents 1 tablespoon is required.
Azure Cosmos DB for Apache Gremlin is a graph database based on Apache TinkerPop to build, a graph computing framework that uses the Gremlin query language.
Content:
- data/ingredient.csv <-- CSV file for storing the ingredient (vertexes).
- data/recipe.csv <-- CSV file for storing the recipe (vertexes).
- data/requires.csv <-- CSV file for storing the relationship between the recipe and ingredient (edges).
- insert_vertices_edges_2.py <-- Version 2 of the Python script to read vertexes and edges CSV data file and insert it into Azure Cosmos DB for Apache Gremlin.
- insert_vertices_edges.py <-- Version 1 of the Python script that hardcodes vertexes and edges to insert into Azure Cosmos DB for Apache Gremlin.
- graph_query.ipynb <-- Sample Notebook to perform interactive Gremlin query over Azure Cosmos DB for Apache Gremlin.
What if we try to extend more (hierarch) vertexes and edges? Graph's unique advantage is high flexibility in data modeling. There are no constraints to adding new data. Just collect and link it.
Enjoy!