Building Subgraphs on Arweave

Reading time: 7 min

In this guide, you will learn how to build and deploy Subgraphs to index the Arweave blockchain.

The Arweave protocol allows developers to store data permanently and that is the main difference between Arweave and IPFS, where IPFS lacks the feature; permanence, and files stored on Arweave can't be changed or deleted.

Arweave already has built numerous libraries for integrating the protocol in a number of different programming languages. For more information you can check:

• Arwiki
• Arweave Resources

The Graph allows you to build custom open APIs called "Subgraphs". Subgraphs are used to tell indexers (server operators) which data to index on a blockchain and save on their servers in order for you to be able to query it at any time using GraphQL.

Graph Node is now able to index data on Arweave protocol. The current integration is only indexing Arweave as a blockchain (blocks and transactions), it is not indexing the stored files yet.

To be able to build and deploy Arweave Subgraphs, you need two packages:

1. @graphprotocol/graph-cli above version 0.30.2 - This is a command-line tool for building and deploying subgraphs. Click here to download using npm.
2. @graphprotocol/graph-ts above version 0.27.0 - This is library of subgraph-specific types. Click here to download using npm.

There are three components of a subgraph:

Defines the data sources of interest, and how they should be processed. Arweave is a new kind of data source.

Here you define which data you want to be able to query after indexing your Subgraph using GraphQL. This is actually similar to a model for an API, where the model defines the structure of a request body.

The requirements for Arweave subgraphs are covered by the existing documentation.

This is the logic that determines how data should be retrieved and stored when someone interacts with the data sources you are listening to. The data gets translated and is stored based off the schema you have listed.

During subgraph development there are two key commands:

$ graph codegen # generates types from the schema file identified in the manifest
$ graph build # generates Web Assembly from the AssemblyScript files, and prepares all the subgraph files in a /build folder

The subgraph manifest subgraph.yaml identifies the data sources for the subgraph, the triggers of interest, and the functions that should be run in response to those triggers. See below for an example subgraph manifest for an Arweave subgraph:

specVersion: 0.0.5
description: Arweave Blocks Indexing
schema:
  file: ./schema.graphql # link to the schema file
dataSources:
  - kind: arweave
    name: arweave-blocks
    network: arweave-mainnet # The Graph only supports Arweave Mainnet
    source:
      owner: 'ID-OF-AN-OWNER' # The public key of an Arweave wallet
      startBlock: 0 # set this to 0 to start indexing from chain genesis
    mapping:
      apiVersion: 0.0.5
      language: wasm/assemblyscript
      file: ./src/blocks.ts # link to the file with the Assemblyscript mappings
      entities:
        - Block
        - Transaction
      blockHandlers:
        - handler: handleBlock # the function name in the mapping file
      transactionHandlers:
        - handler: handleTx # the function name in the mapping file

• Arweave subgraphs introduce a new kind of data source (arweave)
• The network should correspond to a network on the hosting Graph Node. In Subgraph Studio, Arweave's mainnet is arweave-mainnet
• Arweave data sources introduce an optional source.owner field, which is the public key of an Arweave wallet

Arweave data sources support two types of handlers:

• blockHandlers - Run on every new Arweave block. No source.owner is required.
• transactionHandlers - Run on every transaction where the data source's source.owner is the owner. Currently an owner is required for transactionHandlers, if users want to process all transactions they should provide "" as the source.owner

Schema definition describes the structure of the resulting subgraph database and the relationships between entities. This is agnostic of the original data source. There are more details on the subgraph schema definition here.

The handlers for processing events are written in AssemblyScript.

Arweave indexing introduces Arweave-specific data types to the AssemblyScript API.

class Block {
  timestamp: u64
  lastRetarget: u64
  height: u64
  indepHash: Bytes
  nonce: Bytes
  previousBlock: Bytes
  diff: Bytes
  hash: Bytes
  txRoot: Bytes
  txs: Bytes[]
  walletList: Bytes
  rewardAddr: Bytes
  tags: Tag[]
  rewardPool: Bytes
  weaveSize: Bytes
  blockSize: Bytes
  cumulativeDiff: Bytes
  hashListMerkle: Bytes
  poa: ProofOfAccess
}

class Transaction {
  format: u32
  id: Bytes
  lastTx: Bytes
  owner: Bytes
  tags: Tag[]
  target: Bytes
  quantity: Bytes
  data: Bytes
  dataSize: Bytes
  dataRoot: Bytes
  signature: Bytes
  reward: Bytes
}

Block handlers receive a Block, while transactions receive a Transaction.

Writing the mappings of an Arweave Subgraph is very similar to writing the mappings of an Ethereum Subgraph. For more information, click here.

Once your subgraph has been created on your Subgraph Studio dashboard, you can deploy by using the graph deploy CLI command.

graph deploy --access-token <your-access-token>

The GraphQL endpoint for Arweave subgraphs is determined by the schema definition, with the existing API interface. Please visit the GraphQL API documentation for more information.

Here is an example subgraph for reference:

• Example subgraph for Arweave

No, a subgraph can only support data sources from one chain/network.

Currently, The Graph is only indexing Arweave as a blockchain (its blocks and transactions).

This is not currently supported.

The source.owner can be the user's public key or account address.

Data is generally passed into the mappings as Bytes, which if stored directly is returned in the subgraph in a hex format (ex. block and transaction hashes). You may want to convert to a base64 or base64 URL-safe format in your mappings, in order to match what is displayed in block explorers like Arweave Explorer.

The following bytesToBase64(bytes: Uint8Array, urlSafe: boolean): string helper function can be used, and will be added to graph-ts:

const base64Alphabet = [
    "A", "B", "C", "D", "E", "F", "G", "H", "I", "J", "K", "L", "M",
    "N", "O", "P", "Q", "R", "S", "T", "U", "V", "W", "X", "Y", "Z",
    "a", "b", "c", "d", "e", "f", "g", "h", "i", "j", "k", "l", "m",
    "n", "o", "p", "q", "r", "s", "t", "u", "v", "w", "x", "y", "z",
    "0", "1", "2", "3", "4", "5", "6", "7", "8", "9", "+", "/"
];

const base64UrlAlphabet = [
    "A", "B", "C", "D", "E", "F", "G", "H", "I", "J", "K", "L", "M",
    "N", "O", "P", "Q", "R", "S", "T", "U", "V", "W", "X", "Y", "Z",
    "a", "b", "c", "d", "e", "f", "g", "h", "i", "j", "k", "l", "m",
    "n", "o", "p", "q", "r", "s", "t", "u", "v", "w", "x", "y", "z",
    "0", "1", "2", "3", "4", "5", "6", "7", "8", "9", "-", "_"
];

function bytesToBase64(bytes: Uint8Array, urlSafe: boolean): string {
    let alphabet = urlSafe? base64UrlAlphabet : base64Alphabet;

    let result = '', i: i32, l = bytes.length;
    for (i = 2; i < l; i += 3) {
        result += alphabet[bytes[i - 2] >> 2];
        result += alphabet[((bytes[i - 2] & 0x03) << 4) | (bytes[i - 1] >> 4)];
        result += alphabet[((bytes[i - 1] & 0x0F) << 2) | (bytes[i] >> 6)];
        result += alphabet[bytes[i] & 0x3F];
    }
    if (i === l + 1) { // 1 octet yet to write
        result += alphabet[bytes[i - 2] >> 2];
        result += alphabet[(bytes[i - 2] & 0x03) << 4];
        if (!urlSafe) {
            result += "==";
        }
    }
    if (!urlSafe && i === l) { // 2 octets yet to write
        result += alphabet[bytes[i - 2] >> 2];
        result += alphabet[((bytes[i - 2] & 0x03) << 4) | (bytes[i - 1] >> 4)];
        result += alphabet[(bytes[i - 1] & 0x0F) << 2];
        if (!urlSafe) {
            result += "=";
        }
    }
    return result;
}