Skip to main content
Function code is the Python computation invoked by a BlockX Call. You supply positional arguments, read the EVM or table data required by the computation, and return values for the Writer Plugin to handle.

​
Inputs and outputs

For an EVM-derived record, include the entity identifier and any context required by your target schema. Keep integer amounts exact instead of converting them to floating point.

​
Available computation

Use Write EVM data functions for a complete function. The runtime routes backend access through Worker IO.

​
Function identity and version

  • Inline source travels with the input configuration.
  • A registered function is resolved by its identifier.
  • A task pins a function-code snapshot for registered calls and sub-calls.
  • A code snapshot fixes function definitions. Set data-read context separately.
See Function Code View for code loading and snapshot lifecycle.

​
Execution boundaries

  • Produce output records; let the Writer Plugin perform task writes.
  • Use the runtime’s supported modules and capabilities.
  • Keep Calls independent of execution order.
  • Execution limits and audit checks can fail a function before it returns a value.

​
Summary

  • A function connects EVM data reads to derived outputs.
  • Pass historical read context explicitly.
  • Function-code versions and chain-data context are separate.
Continue reading: