Skip to main content

satif_core.code_executors.base

CodeExecutor Objects​

class CodeExecutor(ABC)

Abstract Base Class defining the interface for code execution engines.

This class provides a contract for different strategies of executing Python code, particularly for data transformation tasks. Concrete implementations might execute code locally, in a sandboxed environment, or through other mechanisms.

The primary goal of a CodeExecutor is to:

  1. Accept a description of SDIF data sources.
  2. Establish the necessary database connections and setup based on these sources within its specific execution environment.
  3. Execute a given string of Python code to define a specific function.
  4. Call that function with the established database connection and any extra context.
  5. Return the results from the called function.

execute​

@abstractmethod
def execute(code: str, function_name: str, sdif_sources: Dict[str, Path],
            extra_context: Dict[str, Any]) -> Dict[str, Any]

Abstract method to set up a database environment from SDIF sources, execute a Python code string to define a function, call that function, and return its results.

Concrete implementations will define how the sdif_sources are accessed (e.g., direct file access, copying to a sandbox) and how the database connection is established and managed within their specific environment.

Arguments:

code: A string containing the Python script to be executed. This script is expected to define the function identified by function_name. function_name: The name of the function, defined within the code string, that will be invoked after the code string itself has been executed. sdif_sources: A dictionary mapping desired schema names (strings) to the resolved pathlib.Path objects of the corresponding SDIF database files. The executor is responsible for using these sources to establish the necessary SQLite connection(s) and ATTACH operations before running the user's code. extra_context: A dictionary of additional objects and data to be made available to the transformation logic. Implementations will decide how this context is provided to the executed code.

Returns:

A dictionary, which is the result of successfully calling the user-defined transformation function (function_name). Typically, keys are output filenames and values are the data to be written.

Raises:

satif_core.exceptions.CodeExecutionError: If any error occurs during database setup, code loading, definition, or execution of the code or the subsequent call to function_name. NotImplementedError: If a concrete subclass does not implement this method.