Embedded python interpreter, generate stub source code for autocompletion

Question

I have an application that embeds python and exposes its internal object model as python objects/classes.

For autocompletion/scripting purposes I'd like to extract a mock of the inernal object model, containing the doc tags, structure, functions, etc so I can use it as library source for the IDE autocompletion.

Does someone know of a library, or has some code snippet that could be used to dump those classes to source?

Answer 1

Use the dir() or globals() function to get the list of what has been defined yet. Then, to filter and browse your classes use the inspect module

Example toto.py:

class Example(object):
    """class docstring"""

    def hello(self):
        """hello doctring"""
        pass

Example browse.py:

import inspect
import toto

for name, value in inspect.getmembers(toto):
    # First ignore python defined variables
    if name.startswith('__'):
        continue

    # Now only browse classes
    if not inspect.isclass(value):
        continue
    print "Found class %s with doctring \"%s\"" % (name, inspect.getdoc(value))

    # Only browse functions in the current class
    for sub_name, sub_value in inspect.getmembers(value):
        if not inspect.ismethod(sub_value):
            continue
        print "  Found method %s with docstring \"%s\"" % \
            (sub_name, inspect.getdoc(sub_value))

python browse.py:

Found class Example with doctring "class docstring"
  Found method hello with docstring "hello doctring"

Also, that doesn't really answer your question, but if you're writing a sort of IDE, you can also use the ast module to parse python source files and get information about them

Answer 2

Python data structures are mutable (see What is a monkey patch? ), so extracting a mock would not be enough. You could instead ask the interpreter for possible autocompletion strings dynamically using the dir() built-in function .