Core client module for making JSON-RPC requests to EVM nodes.

This module provides the main client interface for interacting with EVM nodes, handling request creation, response parsing, and client lifecycle management.

Features

• Atomic request ID generation
• Transport abstraction
• Request/response lifecycle management
• Batch request support
• Error handling
• Fluent API for chaining requests

Usage

Basic Usage

# Create a new client
client = Client.new(:http,
  rpc_url: "https://eth-mainnet.example.com",
  timeout: 30_000
)

# Create a raw request
rpc_call = Client.request(client, "eth_blockNumber", [])

# Send the request
{:ok, response} = Client.send(rpc_call)

# Send batch requests
{:ok, responses} =
  client
  |> Client.request("eth_blockNumber", [])
  |> Client.request("eth_getBalance", [address, block])
  |> Client.send()

Using Raw Requests

# Create a raw request
request = Request.new("eth_blockNumber", [], 1)

# Send the request
{:ok, response} = Client.send(request, client)
# or
{:ok, response} = Client.send(client, request)

# Send batch requests
{:ok, responses} = Client.send([request1, request2], client)
# or
{:ok, responses} = Client.send(client, [request1, request2])

Client Configuration

The client accepts the following options:

• :rpc_url - (Required) The endpoint URL
• other options that are specific to the transport type

Request ID Generation

The client uses Erlang's :atomics for thread-safe, monotonic request ID generation. This ensures:

• Unique IDs across concurrent requests
• No ID collisions in batch requests
• Efficient ID allocation
• Process-independent ID tracking

Transport Layer

The client supports different transport mechanisms through the Exth.Transport.Transportable protocol:

• Built-in HTTP transport using Tesla/Mint
• Custom transport implementations
• Future support for WebSocket and IPC

Error Handling

The client provides consistent error handling:

• {:ok, response} - Successful request
• {:error, reason} - Request failed

Best Practices

• Reuse client instances when possible
• Use batch requests for multiple calls
• Implement appropriate timeouts
• Handle errors gracefully
• Monitor client health
• Clean up resources when done

See Exth.Transport for transport details.