steer-protocol-watcher-ts

Source

• Subgraph: filecoin-subgraph
  • NOTE: Generating watcher using codegen config works without any changes in subgraph repo

Setup

• Run the following command to install required packages:

  yarn
  

• Run build:

  yarn build
  

• Create a postgres12 database for the watcher:

  sudo su - postgres
  createdb steer-protocol-watcher-ts
  

• If the watcher is an active watcher:

  Create database for the job queue and enable the pgcrypto extension on them (https://github.com/timgit/pg-boss/blob/master/docs/usage.md#intro):

  createdb steer-protocol-watcher-ts-job-queue
  

  postgres@tesla:~$ psql -U postgres -h localhost steer-protocol-watcher-ts-job-queue
  Password for user postgres:
  psql (12.7 (Ubuntu 12.7-1.pgdg18.04+1))
  SSL connection (protocol: TLSv1.3, cipher: TLS_AES_256_GCM_SHA384, bits: 256, compression: off)
  Type "help" for help.
  
  steer-protocol-watcher-ts-job-queue=# CREATE EXTENSION pgcrypto;
  CREATE EXTENSION
  steer-protocol-watcher-ts-job-queue=# exit
  

• In the config file:

  • Update the database connection settings.

  • Update the upstream config and provide the ipld-eth-server GQL API endpoint.

  • Update the server config with state checkpoint settings.

Customize

• Indexing on an event:

  • Edit the custom hook function handleEvent (triggered on an event) in hooks.ts to perform corresponding indexing using the Indexer object.

  • While using the indexer storage methods for indexing, pass diff as true if default state is desired to be generated using the state variables being indexed.

• Generating state:

  • Edit the custom hook function createInitialState (triggered if the watcher passes the start block, checkpoint: true) in hooks.ts to save an initial State using the Indexer object.

  • Edit the custom hook function createStateDiff (triggered on a block) in hooks.ts to save the state in a diff State using the Indexer object. The default state (if exists) is updated.

  • Edit the custom hook function createStateCheckpoint (triggered just before default and CLI checkpoint) in hooks.ts to save the state in a checkpoint State using the Indexer object.

GQL Caching

To enable GQL requests caching:

• Update the server.gql.cache config with required settings.

• In the GQL schema file, use the cacheControl directive to apply cache hints at schema level.

  • Eg. Set inheritMaxAge to true for non-scalar fields of a type.

• In the GQL resolvers file, uncomment the setGQLCacheHints() calls in resolvers for required queries.

Run

• If the watcher is a lazy watcher:

  • Run the server:

    yarn server
    

    GQL console: http://localhost:3008/graphql

• If the watcher is an active watcher:

  • Run the job-runner:

    yarn job-runner
    

  • Run the server:

    yarn server
    

    GQL console: http://localhost:3008/graphql

  • To watch a contract:

    yarn watch:contract --address <contract-address> --kind <contract-kind> --checkpoint <true | false> --starting-block [block-number]
    

    • address: Address or identifier of the contract to be watched.
    • kind: Kind of the contract.
    • checkpoint: Turn checkpointing on (true | false).
    • starting-block: Starting block for the contract (default: 1).

    Examples:

    Watch a contract with its address and checkpointing on:

    yarn watch:contract --address 0x1F78641644feB8b64642e833cE4AFE93DD6e7833 --kind ERC20 --checkpoint true
    

    Watch a contract with its identifier and checkpointing on:

    yarn watch:contract --address MyProtocol --kind protocol --checkpoint true
    

  • To fill a block range:

    yarn fill --start-block <from-block> --end-block <to-block>
    

    • start-block: Block number to start filling from.
    • end-block: Block number till which to fill.

  • To create a checkpoint for a contract:

    yarn checkpoint create --address <contract-address> --block-hash [block-hash]
    

    • address: Address or identifier of the contract for which to create a checkpoint.
    • block-hash: Hash of a block (in the pruned region) at which to create the checkpoint (default: latest canonical block hash).

  • To verify a checkpoint:

    yarn checkpoint verify --cid <checkpoint-cid>
    

    cid: CID of the checkpoint for which to verify.

  • To reset the watcher to a previous block number:

    • Reset watcher:

      yarn reset watcher --block-number <previous-block-number>
      

    • Reset job-queue:

      yarn reset job-queue
      

    • Reset state:

      yarn reset state --block-number <previous-block-number>
      

    • block-number: Block number to which to reset the watcher.

  • To export and import the watcher state:

    • In source watcher, export watcher state:

      yarn export-state --export-file [export-file-path] --block-number [snapshot-block-height]
      

      • export-file: Path of file to which to export the watcher data.
      • block-number: Block height at which to take snapshot for export.

    • In target watcher, run job-runner:

      yarn job-runner
      

    • Import watcher state:

      yarn import-state --import-file <import-file-path>
      

      • import-file: Path of file from which to import the watcher data.

    • Run server:

      yarn server
      

  • To inspect a CID:

    yarn inspect-cid --cid <cid>
    

    • cid: CID to be inspected.