Skip to content

feat(py): added parsing for Python definitions

Jonathan Shobrookrequested to merge
python-definitions into main

What does this MR do?

This MR implements definition extraction for Python. It builds off the analyzer pattern established by the Ruby definitions MR.

Given a Python file, the parser will return a list of DefinitionInfo objects. Each will contain the following:

  • The definition type (class, function, or lambda)
  • Name of the function
  • Fully qualified name (FQN) of the definition, with metadata, according to the new schema
  • Match information from the ast-grep rule engine
  • All captured ast-grep variables

Notably, there are a few limitations:

  • We should compute aliases for each iteration, as described here
    • Saving for a future iteration
  • Anonymous functions (i.e. unnamed lambdas) are intentionally not handled here because they can't be referenced
  • We do not capture nested lambda definitions (e.g. funcs = {"fn_1": lambda x: x}) even though they are technically named
    • Saving for a future iteration
  • Scope changes using nonlocal or global keywords are not handled
  • Dynamic classes are not captured (e.g. DynamicClass = type("DynamicClass", (object,), {"method": lambda self: "Hello"}))
  • Any non-callable definition (e.g. variable assignments, types, etc.) are not captured

Related issues

Edited by Jonathan Shobrook

Merge request reports