Files
Phoenix/sphinxtools/modulehunter.py

646 lines
18 KiB
Python

# -*- coding: utf-8 -*-
# Describe classes, methods and functions in a module.
# Works with user-defined modules, all Python library
# modules, including built-in modules.
import os
import sys
import glob
import types
import cPickle
import imp
import traceback
import pkgutil
import __builtin__
from inspect import getargspec, ismodule, getdoc, getmodule, getcomments, isfunction
from inspect import ismethoddescriptor, getsource, ismemberdescriptor, isgetsetdescriptor
from inspect import isbuiltin, isclass, getfile, ismethod
from librarydescription import Library, Module, Class
from librarydescription import Method, Property, Attribute
import inheritance
from constants import object_types, EXCLUDED_ATTRS, MODULE_TO_ICON
from constants import CONSTANT_RE
reload(sys)
sys.setdefaultencoding("utf-8")
try:
import wx
except ImportError:
wxPath = ''
basePath = ''
for path in sys.path:
if 'wx-' in path:
dummy, newPath = os.path.split(path)
if newPath > wxPath:
wxPath = newPath
basePath = dummy
if not wxPath:
raise Exception('Unable to find the wx package')
sys.path.insert(0, os.path.join(basePath, wxPath))
import wx
print '\nUSING VERSION: %s\n'%wx.VERSION_STRING
if hasattr(os.path, "relpath"):
relpath = os.path.relpath # since Python 2.6
else:
def relpath(path, start=os.path.curdir):
"""Return a relative version of a path"""
if not path:
raise ValueError("no path specified")
start_list = os.path.abspath(start).split(os.path.sep)
path_list = os.path.abspath(path).split(os.path.sep)
# Work out how much of the filepath is shared by start and path.
i = len(os.path.commonprefix([start_list, path_list]))
rel_list = [os.path.pardir] * (len(start_list)-i) + path_list[i:]
if not rel_list:
return os.path.curdir
return os.path.join(*rel_list)
def format_traceback():
t, v = sys.exc_info()[:2]
message = "".join(traceback.format_exception_only(t, v)).replace("\n", " ")
return message.strip()
def format_method(method):
method = method.strip()
if "def " not in method:
return None, None
indx1, indx2, indx3 = method.index("def "), method.index("("), method.rindex(")")
name = method[indx1+4:indx2]
signature = method[indx2+1:indx3]
if "\n" in signature:
sig = signature.split("\n")
params = ""
for s in sig:
params += s.strip() + " "
params = params.rstrip()
else:
params = signature
return name, params
def analyze_params(obj, signature):
params = signature.split(",")
if "self" in params:
params.remove("self")
signature = ",".join(params)
param_tuple = []
if not params:
return signature, param_tuple
try:
arginfo = getargspec(obj)
except TypeError:
arginfo = None
pevals = {}
if arginfo:
args = arginfo[0]
argsvar = arginfo[1]
if arginfo[3]:
dl = len(arginfo[3])
al = len(args)
defargs = args[al-dl:al]
info = arginfo[3]
for d, i in zip(defargs, info):
pevals[d] = i
for par in params:
p = par.strip()
pvalue = peval = None
if "=" in p:
all_values = p.split("=")
pname, pvalue = all_values[0].strip(), "=".join(all_values[1:]).strip()
pvalue = pvalue.strip()
if pname in pevals:
try:
peval = unicode(pevals[pname])
except UnicodeDecodeError:
peval = repr(pevals[pname])
except TypeError:
peval = u''
else:
pname = p
param_tuple.append((pname, pvalue, peval))
return signature, param_tuple
def get_constructor(source):
description = ""
hasComma = False
for line in source.split("\n"):
if '#' in line:
line = line[0:line.index('#')].strip()
if ":" in line:
hasComma = True
commaPos = line.index(":")
if ('"""' in line or "'''" in line) and hasComma:
break
if hasComma and ' def ' in line:
defPos = line.index(' def ')
if defPos > commaPos:
break
description += " " + line.strip()
if "):" in line or ") :" in line or ") :" in line:
break
return description
def inspect_source(method_class, obj, source):
description = get_constructor(source)
name, params = format_method(description)
if name is None:
return
signature, param_tuple = analyze_params(obj, params)
method_class.arguments = param_tuple
method_class.signature = description.strip()
if "classmethod " in description or is_classmethod(obj):
method_class.kind = object_types.CLASS_METHOD
elif "staticmethod " in description:
method_class.kind = object_types.STATIC_METHOD
def is_classmethod(instancemethod):
" Determine if an instancemethod is a classmethod. "
if hasattr(instancemethod, 'im_self'):
return instancemethod.im_self is not None
return False
def describe_func(obj, parent_class, module_name):
"""
Describe the function object passed as argument.
If this is a method object, the second argument will
be passed as True.
"""
try:
name = obj.__name__
except AttributeError:
# Funny comtypes...
return
if name.startswith("_") and "__init__" not in name:
return
name = parent_class.name + "." + name
docs = getdoc(obj)
comments = getcomments(obj)
if isfunction(obj):
method = object_types.FUNCTION
elif ismethod(obj):
method = object_types.METHOD
elif ismethoddescriptor(obj):
method = object_types.METHOD_DESCRIPTOR
if isinstance(obj, types.MethodType):
method = object_types.INSTANCE_METHOD
try:
source_code = getsource(obj)
except (IOError, TypeError):
source_code = ""
klass = Method(name, method)
klass.docs = docs
klass_module = getmodule(obj)
if klass_module and klass_module.__name__ != module_name:
klass.is_redundant = True
if source_code:
inspect_source(klass, obj, source_code)
klass.number_lines = "%d"%len(source_code.split("\n"))
if isinstance(obj, staticmethod):
klass.method = method = object_types.STATIC_METHOD
try:
if method in [object_types.METHOD, object_types.METHOD_DESCRIPTOR, object_types.INSTANCE_METHOD]:
code = obj.im_func.func_code
elif method == object_types.STATIC_METHOD:
code = obj.im_func.func_code
else:
code = obj.func_code
except AttributeError:
code = None
if code is not None:
klass.firstlineno = "%d"%code.co_firstlineno
parent_class.Add(klass)
def describe_class(obj, module_class, module_name, constants):
"""
Describe the class object passed as argument,
including its methods.
"""
class_name = obj.__name__
if class_name == "object":
return
class_name = module_class.name + "." + class_name
docs = getdoc(obj)
comments = getcomments(obj)
obj_dict = obj.__dict__
klass = Class(class_name, obj)
count = 0
for name in obj_dict:
if name.startswith("_") and "__init__" not in name:
continue
if name in EXCLUDED_ATTRS:
continue
try:
item = getattr(obj, name)
except AttributeError:
# Thanks to ReportLab for this funny exception...
continue
if ismodule(item):
continue
if ismemberdescriptor(item) or isgetsetdescriptor(item):
continue
if isbuiltin(item):
count += 1
elif ismethod(item) or isfunction(item) or ismethoddescriptor(item) or \
isinstance(item, types.MethodType):
count += 1
describe_func(item, klass, module_name)
elif isclass(item):
count += 1
describe_class(item, klass, module_name, constants)
else:
name = class_name + "." + name
if isinstance(item, property):
item_class = Property(name, item)
klass.Add(item_class)
item_module = getmodule(obj)
if item_module and item_module.__name__ != module_name:
item_class.is_redundant = True
else:
item_class = Attribute(name, type(item), item)
klass.Add(item_class)
if constants:
item_class.is_redundant = name not in constants
count += 1
klass.docs = docs
klass.comments = comments
klass_module = getmodule(obj)
if klass_module and klass_module.__name__ != module_name:
klass.is_redundant = True
else:
klass.inheritance_diagram = inheritance.InheritanceDiagram([obj], klass)
module_class.Add(klass)
try:
source_code = getsource(obj)
except (IOError, TypeError):
source_code = ""
if source_code:
description = get_constructor(source_code)
if '(' not in description and ':' in description:
description = description[0:description.index(':')]
klass.signature = description.strip()
klass.number_lines = '%d'%len(source_code.split("\n"))
def describe_module(module, kind, constants=[]):
"""
Describe the module object passed as argument
including its classes and functions.
"""
module_name = module.__name__
if kind == object_types.LIBRARY:
klass = Library(module_name)
else:
klass = Module(module_name, kind)
klass.docs = getdoc(module)
klass.comments = getcomments(module)
klass.filename = module.__file__
inheritance_diagram = []
count = 0
for name in dir(module):
if name in EXCLUDED_ATTRS:
continue
obj = getattr(module, name)
if ismodule(obj):
continue
if ismemberdescriptor(obj) or isgetsetdescriptor(obj):
continue
if isclass(obj):
count += 1
describe_class(obj, klass, module_name, constants)
if obj.__module__ == module.__name__:
inheritance_diagram.append(obj)
elif isbuiltin(obj):
count += 1
elif ismethod(obj) or isfunction(obj) or ismethoddescriptor(obj) or \
isinstance(obj, types.MethodType):
count +=1
describe_func(obj, klass, module_name)
else:
attribute = Attribute(module_name + "." + name, type(obj), obj)
klass.Add(attribute)
if constants:
attribute.is_redundant = name not in constants
if kind not in [object_types.PACKAGE, object_types.LIBRARY]:
if inheritance_diagram and len(inheritance_diagram) < 20:
klass.inheritance_diagram = inheritance.InheritanceDiagram(inheritance_diagram, klass)
return klass, count
def Import(init_name, import_name, full_process=True):
directory, module_name = os.path.split(init_name)
dirname = os.path.dirname(directory)
if not full_process:
path = list(sys.path)
sys.path.insert(0, dirname)
f = None
try:
f, filename, description = imp.find_module(import_name, [dirname])
mainmod = imp.load_module(import_name, f, filename, description)
except (ImportError, NameError):
message = format_traceback()
message += ' Please check that the "Import command" text box is correctly filled'
print "Error: %s"%message
if not full_process:
sys.path = path[:]
return
if f:
f.close()
if not full_process:
sys.path = path[:]
try:
version = mainmod.__version__
except AttributeError:
try:
version = mainmod.__VERSION__
except AttributeError:
print "Warning: Library '%s' has no __version__ or __VERSION__ attribute. Please specify it in the 'Import command' textbox"%import_name
return
print version
return mainmod
def PrintProgress(name, looped_names):
looped_names.append(name)
if len(looped_names) == 5:
message = ", ".join(looped_names)
looped_names = []
print message
return looped_names
def FindModuleType(filename):
splitext = os.path.splitext(filename)[0]
for extension, icon, description in MODULE_TO_ICON:
if os.path.isfile(splitext + extension):
return icon
def SubImport(import_string, module, parent_class, ispkg):
try:
submod = __import__(import_string, fromlist=[module])
except:
# pubsub and Editra can be funny sometimes...
message = "Unable to import module/package '%s.%s'.\n Exception was: %s"%(import_string, module, format_traceback())
print "\nWARNING: %s\n"%message
return None, 0
if not ismodule(submod):
return None, 0
filename = getfile(submod)
subpath = os.path.dirname(filename)
if subpath not in sys.path:
sys.path.append(subpath)
if ispkg:
kind = object_types.PACKAGE
else:
kind = FindModuleType(filename)
constants = []
if kind in [object_types.PY_MODULE, object_types.PACKAGE]:
contents = open(filename, "rt").read()
consts = CONSTANT_RE.findall(contents)
for c in consts:
if "," in c:
c = c.split(",")
constants.extend([v.strip() for v in c])
else:
constants.append(c.strip())
module_class, count = describe_module(submod, kind=kind, constants=constants)
parent_class.Add(module_class)
return module_class, count
def ModuleHunter(init_name, import_name, version):
pickle_file = os.path.join(os.getcwd(), 'docs', 'sphinx', 'wx%s.pkl'%import_name)
if os.path.isfile(pickle_file):
fid = open(pickle_file, 'rb')
library_class = cPickle.load(fid)
fid.close()
library_class.Walk(library_class)
return
path = list(sys.path)
directory, module_name = os.path.split(init_name)
path = list(sys.path)
sys.path.insert(0, os.path.dirname(directory))
mainmod = Import(init_name, import_name)
if mainmod is None:
return
message = "Importing main library '%s'..."%import_name
print "Message: %s"%message
module_name = os.path.splitext(getfile(mainmod))[0] + ".py"
contents = open(module_name, "rt").read()
constants = CONSTANT_RE.findall(contents)
library_class, count = describe_module(mainmod, kind=object_types.LIBRARY, constants=constants)
library_class.name = "%s-%s"%(import_name, version)
message = "Main library '%s' imported..."%library_class.name
print "Message: %s"%message
message = "Importing sub-modules and sub-packages...\n"
print "Message: %s"%message
looped_names = []
ancestors_dict = {import_name: library_class}
for importer, module_name, ispkg in pkgutil.walk_packages(path=[directory],
prefix=import_name+".",
onerror=lambda x: None):
import_string = module_name
splitted = module_name.split(".")
fromlist = splitted[-1]
parent_name = ".".join(splitted[0:-1])
parent_class = ancestors_dict[parent_name]
module_class, count = SubImport(import_string, fromlist, parent_class, ispkg)
if module_class is None:
continue
looped_names = PrintProgress(module_name, looped_names)
if module_name not in ancestors_dict:
ancestors_dict[module_name] = module_class
major, minor, micro, release = sys.version_info[0:-1]
pythonVersion = u"%d.%d.%d-%s"%(major, minor, micro, release)
library_class.python_version = pythonVersion
library_class.Save()
sys.path[:] = path # restore
fid = open(pickle_file, 'wb')
cPickle.dump(library_class, fid)
fid.close()
library_class.Walk(library_class)
if __name__ == "__main__":
argv = sys.argv[1:]
if len(argv) == 2:
init_name, import_name = argv
Import(init_name, import_name, full_process=False)
else:
init_name, import_name, version, save_dir = argv
ModuleHunter(init_name, import_name, version, save_dir)