pulsarprotos

This is the API schema for the Pulsar RPC. The schema here describes how RPC calls are shaped and function using Protocol Buffer definitions. Having a unified abstraction helps us to generate client and server code across different languages so that we remain mostly independent of transport layer specifics.

Transfers

Pulsar allows to search for paginated token transfers associated to the specified wallet addresses.

{
  "filter": {
    "paging": {
      "kind": "page",
      "start": "0",
      "stop": "49"
    }
  },
  "search": [
    {
      "address": "0x4838...5f97"
    }
  ]
}

{
  "result": [
    {
      "address": "0x4838...5f97",
      "transfers": [
        {
          "amount": "420000000000000000000",
          "created_at": "1689001255",
          "id": "a9526eca-953e-49f7-aa4c-12fc06132124",
          "to": "0x4838...5f97"
        }
      ]
    }
  ]
}

Limitations

All RPCs implement a hard limit on the maximum amount of actions that can be processed within the same request. Specifically no more than 100 actions can be received on any given endpoint in a single call. Further note that the hard cap of returned objects is 1000. No more than 1000 objects can be received from any endpoint in a single call.

Pagination

Pagination works by providing paging pointers for search query objects. The pointers start and stop are always inclusive. In the request example below the response would contain the first batch of 50 items. Depending on the requested resources, consecutive calls with the exact same pointers may not result in the same objects received, given the dynamic nature of the underlying system.

{
  "filter": {
    "paging": {
      "kind": "page",
      "start": "0",
      "stop": "49"
    }
  }
}

Providing paging pointers in requests containing multiple search query objects causes the system to apply paging to each search query result. Consider resources to be searched from two lists using two search query objects. Say the lists look like [A B C D] and [E F G H]. Searching for both results from both lists in a single call using start=0 and stop=1 would return [A B] and [E F]. If paging is desired for large lists of objects, single search query objects may be more applicable, depending on the use case.

There is only one kind of paging pointer supported at the moment, kind=page. Kind "page" is the default and does not have to be provided. Using kind "page" may result in less received objects than requested, but never more.

Development

Merging changes into the main branch triggers several Github Actions to auto generate code for different programming languages. Once the package that you are interested in is synchronized, you can import your updated dependency.

• .github/workflows/pbf-go.yaml updates https://github.com/0xSplits/pulsargocode
• .github/workflows/pbf-ts.yaml updates https://github.com/0xSplits/pulsartscode

Formatting

You need to install the Clang Formatter on your system in order to format the .proto files properly. If you are using VS Code you can also install the VS Code Extension and format files on save.

brew install clang-format

clang-format -i $(find ./pbf -name "*.proto")