The Starter Package is an example that you can use as a starting point for developing your SubQuery project. A SubQuery package defines which data The SubQuery will index from the Substrate blockchain, and how it will store it.

• Typescript are required to compile project and define types.

• Both SubQuery CLI and generated Project have dependencies and require Node.

Install SubQuery CLI globally on your terminal by using NPM:

npm install -g @subql/cli

Run help to see available commands and usage provide by CLI

subql help

Inside the directory in which you want to create the SubQuery project, simply replace project-name with your project name and run the command:

subql init --starter project-name

Then you should see a folder with your project name has been created inside the directory, you can use this as the start point of your project. And the files should be identical as in the Directory Structure.

Last, under the project directory, run following command to install all the dependency.

yarn install

In the starter package, we have provided a simple example of project configuration. You will be mainly working on the following files:

• The Manifest in project.yaml
• The GraphQL Schema in schema.graphql
• The Mapping functions in src/mappings/ directory

For more information on how to write the SubQuery, check out our doc section on Define the SubQuery

In order to index your SubQuery project, it is mandatory to build your project first. Run this command under the project directory.

yarn codegen

In order to deploy your SubQuery project to our hosted service, it is mandatory to pack your configuration before upload. Run pack command from root directory of your project will automatically generate a your-project-name.tgz file.

yarn build

Under the project directory run following command:

docker-compose pull && docker-compose up

Open your browser and head to http://localhost:3000.

Finally, you should see a GraphQL playground is showing in the explorer and the schemas that ready to query.

For the subql-starter project, you can try to query with the following code to get a taste of how it works.

{
  query {
    starterEntities(first: 10) {
      nodes {
        field1
        field2
        field3
      }
    }
  }
}

We built an extension to the subquery service that automatically created graphQL types and mapping handlers for all encointer events. The graphQL types will look as follows:

type Transferred @entity {
        id: ID!
        blockHeight: BigInt! @index
        timestamp: BigInt! @index
        arg0: String! @index
        arg1: String! @index
        arg2: String! @index
        arg3: Float!
}

Where arg0, arg1.. are simply the values of the arguments of the event. There are no extra steps that need to be taken in order to generate those types. It is inclused in the yarn codegen command.

There is some extra configuration than can be made when generating the graphQL schema.

Per default the arguments of the event will simply be parsed to a string using the .toString() method. Should there be a need to do some extra conversion between a Substrate type and the corresponding graphQL type this conversion can be defined in src/codeGen/typeConversions.ts by adding an entry to the conversions object, like in the example below

const conversions = {
    CommunityIdentifier: {
        graphQlTypeName: "String",
        convert: (input) => {
            let inp = input.toHuman();
            const geohash = inp["geohash"];
            const digest = inp["digest"];
            let buffer;
            if (digest.startsWith("0x")) {
                buffer = Buffer.from(input.toHuman()["digest"].slice(2), "hex");
            } else {
                buffer = Buffer.from(input.toHuman()["digest"], "utf-8");
            }

            return geohash + bs58.encode(Uint8Array.from(buffer));
        },
    },
};

The above example defines a conversion from the Encointer type CommunityIdentifier to the graphQL type String using a conversion function pasring the cid and converting to Encointer's custom Base58 cid format.

In src/codeGen/config.ts there is a indexedTypes array. Event arguments of types that are contained in this array will receive an @index in the corresponding GraphQl schema for faster queries. For example T::AccountId is contained in the array, because we need to be able to effiecliently run queries like finding all transactions of a specific account.

Also in src/codeGen/config.ts there is an object called shortEventNameMap. This is to map long event names to shorter ones as unfortunately in Postgres the length of a relation name is limited to 63 chars and with long event names this limit will be exeeded for constraint names, leading to trunctation of the relation name, which will lead to errors about duplicate relation names.