Server folder
contains everything to run the Apollo GraphQL server on neo4j
- npm run database creates new Neo4J in docker container and starts it
- npm start runs the GraphQL server
Starting the Neo4J database
For creating the neo4j database for the first time, use npm run database on the server folder. This command automatically starts the container.
cd server
npm install
npm run database
npm run database creates a new container and starts it. In case you have stopped the container, start the neo4j container by running docker start neo4j.
Starting the GraphQL server
npm start on the server folder starts the GraphQL server. Keep it running.
cd server
npm start
The GraphQL data can viewed using http://localhost:4001/graphql. The tool has DOCS for writing queries and mutations.
Running Cipher queries
To open a interactive docker shell, use the following command (where neo4j is the name of the container). User name is neo4j and password is test.
docker exec --interactive --tty neo4j bin/cypher-shell
server/database.cql has all the CQL queries that you can run. Run those queries in the interactive terminal. Type :exit to exit the shell.
Client folder
contains code to seed, query, mutate, export database using Apollo client
Keep the server running in a separate terminal window. And create a new terminal window where the following client commands are run.
cd client
npm install
npm run seed
- npm run seed creates all products from the api
- npm run find queries for products matching tag, size
- npm run update toggles the availability of a variant
- npm run delete deletes product by handle
- npm run export exports data from neo4j and compares it with data from api
- npm run clear clears all data in neo4j database
Syntax for find
npm run find <tag> <size>
To supply only size, type null for tag. Size is case-insensitive, use XS or xs. Tag is case sensitive. For example, use "Style_Removable Padded" for tag in argument. Enclose the tag with single or double quote. npm run find "Style_Removable Padded" or npm run find null s or npm run find null S or npm run find "Style_Removable Padded" s are all valid.
Syntax for update
npm run update <handle> <size>
Argument handle is the product handle. Argument size is the variant size (option1s). Both arguments are mandatory.
For example, npm run update pure-eyes-falbala-bikini-top s
Syntax for delete
npm run delete <handle>
Argument handle is the product handle is mandatory. For example, npm run delete pure-eyes-falbala-bikini-top
Syntax for export
npm run export
This exports data from neo4j database and compares it with data from api. No major differences were found. The only minor difference was date formatting. But there is no loss in data over there. In some places, date from API is formatted as "2019-05-28T06:00:55+00:00". The exported date is "2019-05-28T06:00:55Z" with no loss in precision. Both are UTC times with slightly different formatting.
Additional notes
Searching on array of strings
Searching on array of strings is not possible with neo4j-graphql-js built-in filters. So, I converted the array of strings
into a comma separated string. Using the following filter works: { tags_contains: $tag }, where $tag is prefixed and suffixed with commas. An example is found in find.js. Note that the separator for tags can be any set of arbitrary string, not necessarily a comma.
Case-insensitive search
To perform case-insensitive search, another field should be created in GraphQL Schema which are all lowercase characters. For eg, if option1 is the field on which you want to perform case-insensitive search, create a shadow field, option1s which has the same string in lowercase. Take the input string and convert it to lowercase, and do a search on option1s instead of option1. An example is found in find.js. This becomes painful if you want it to be a general feature in your app.
Unique constraints
ID in GraphQL Schema does not enforce unique constraints in Neo4J database. Please create a unique constraint using Cipher Query on Neo4J database wherever applicable.
CREATE CONSTRAINT ON (product:Product) ASSERT product.id IS UNIQUE;
Similar CQL queries exist for indexes.
Deleting a node using CQL vs DeleteXXX mutation
Deleting a node using CQL does not always work.
MATCH(p:Product { id: "1826623357018" }) DELETE p;
The above delete statement gives the following error: Cannot delete node<4689>, because it still has relationships. To delete this node, you must first delete its relationships.
However, the DeleteProduct mutation from the augmented schema not only deletes the product but drops all the relations. That is how npm run clear command on the client folder deletes the entire database by performing DeleteXXX mutations.
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent