example_of_suggest_function() - Code Metrics - Inspection of "Handle cases where traceback is None" - SylvainDe/DidYouMean-Python - Measure and Improve Code Quality continuously with Scrutinizer

Completed
Push — master ( 84ff32...ae1382 )

by De
created 2016-10-06 09:08 UTC
example_of_suggest_function() A

↳ Parent: Project
Complexity

Conditions
Size

Total Lines
Duplication

Lines	0
Ratio	0 %
Importance

Changes	1
Bugs	0	Features	0
Metric	Value
c	1
b	0
f	0
dl	0
loc	18
rs	9.4285
cc	1
# -*- coding: utf-8
"""Logic to add suggestions to exceptions."""
import keyword
import difflib
import didyoumean_re as re
import itertools
import inspect
import errno
import os
from collections import namedtuple


#: Standard modules we'll consider while searching for undefined values
# To be completed
# Potential candidates :
#  - sys.builtin_module_names
# https://docs.python.org/2/library/sys.html#sys.builtin_module_names
#  - sys.modules
# https://docs.python.org/2/library/sys.html#sys.modules
#  - pkgutil.iter_modules
# https://docs.python.org/2/library/pkgutil.html#pkgutil.iter_modules
STAND_MODULES = set(['string', 'os', 'sys', 're', 'math', 'random',
                     'datetime', 'timeit', 'unittest', 'itertools',
                     'functools', 'collections', '__future__'])

#: Almost synonyms methods that can be confused from one type to another
# To be completed
SYNONYMS_SETS = [set(['add', 'append']), set(['extend', 'update'])]

#: Maximum number of files suggested
MAX_NB_FILES = 4


# Helper function for string manipulation
def quote(string):
    """Surround string with single quotes."""
    return "'{0}'".format(string)


def get_close_matches(word, possibilities):
    """
    Return a list of the best "good enough" matches.

    Wrapper around difflib.get_close_matches() to be able to
    change default values or implementation details easily.
    """
    return difflib.get_close_matches(word, possibilities, 3, 0.7)


def get_suggestion_string(sugg):
    """Return the suggestion list as a string."""
    sugg = list(sugg)
    return ". Did you mean " + ", ".join(sugg) + "?" if sugg else ""


# Helper functions for code introspection
def get_subclasses(klass):
    """Get the subclasses of a class.

    Get the set of direct/indirect subclasses of a class including itself.
    """
    if hasattr(klass, '__subclasses__'):
        try:
            subclasses = set(klass.__subclasses__())
        except TypeError:
            try:
                subclasses = set(klass.__subclasses__(klass))
            except TypeError:
                subclasses = set()
    else:
        subclasses = set()
    for derived in set(subclasses):
        subclasses.update(get_subclasses(derived))
    subclasses.add(klass)
    return subclasses


def get_types_for_str_using_inheritance(name):
    """Get types corresponding to a string name.

    This goes through all defined classes. Therefore, it :
    - does not include old style classes on Python 2.x
    - is to be called as late as possible to ensure wanted type is defined.
    """
    return set(c for c in get_subclasses(object) if c.__name__ == name)


def get_types_for_str_using_names(name, frame):
    """Get types corresponding to a string name using names in frame.

    This does not find everything as builtin types for instance may not
    be in the names.
    """
    return set(obj
               for obj, _ in get_objects_in_frame(frame).get(name, [])
               if inspect.isclass(obj) and obj.__name__ == name)


def get_types_for_str(tp_name, frame):
    """Get a list of candidate types from a string.

    String corresponds to the tp_name as described in :
    https://docs.python.org/2/c-api/typeobj.html#c.PyTypeObject.tp_name
    as it is the name used in exception messages. It may include full path
    with module, subpackage, package but this is just removed in current
    implementation to search only based on the type name.

    Lookup uses both class hierarchy and name lookup as the first may miss
    old style classes on Python 2 and second does find them.
    Just like get_types_for_str_using_inheritance, this needs to be called
    as late as possible but because it requires a frame, there is not much
    choice anyway.
    """
    name = tp_name.split('.')[-1]
    res = set.union(
        get_types_for_str_using_inheritance(name),
        get_types_for_str_using_names(name, frame))
    assert all(inspect.isclass(t) and t.__name__ == name for t in res)
    return res


def merge_dict(*dicts):
    """Merge dicts and return a dictionnary mapping key to list of values.

    Order of the values corresponds to the order of the original dicts.
    """
    ret = dict()
    for dict_ in dicts:
        for key, val in dict_.items():
            ret.setdefault(key, []).append(val)
    return ret

ScopedObj = namedtuple('ScopedObj', 'obj scope')


def add_scope_to_dict(dict_, scope):
    """Convert name:obj dict to name:ScopedObj(obj,scope) dict."""
    return dict((k, ScopedObj(v, scope)) for k, v in dict_.items())


def get_objects_in_frame(frame):
    """Get objects defined in a given frame.

    This includes variable, types, builtins, etc.
    """
    # https://www.python.org/dev/peps/pep-0227/ PEP227 Statically Nested Scopes
    # "Under this proposal, it will not be possible to gain dictionary-style
    #      access to all visible scopes."
    # https://www.python.org/dev/peps/pep-3104/ PEP 3104 Access to Names in
    #      Outer Scopes
    return merge_dict(  # LEGB Rule (missing E atm - not sure if a problem)
        add_scope_to_dict(frame.f_locals, 'local'),
        add_scope_to_dict(frame.f_globals, 'global'),
        add_scope_to_dict(frame.f_builtins, 'builtin'),
    )


def import_from_frame(module_name, frame):
    """Wrapper around import to use information from frame."""
    if frame is None:
        return None
    return __import__(
        module_name,
        frame.f_globals,
        frame.f_locals)


# Functions related to NameError
def suggest_name_not_defined(value, frame, groups):
    """Get the suggestions for name in case of NameError."""
    del value  # unused param
    name, = groups
    objs = get_objects_in_frame(frame)
    return itertools.chain(
        suggest_name_as_attribute(name, objs),
        suggest_name_as_standard_module(name),
        suggest_name_as_name_typo(name, objs),
        suggest_name_as_keyword_typo(name),
        suggest_name_as_missing_import(name, objs, frame),
        suggest_name_as_special_case(name))


def suggest_name_as_attribute(name, objdict):
    """Suggest that name could be an attribute of an object.

    Example: 'do_stuff()' -> 'self.do_stuff()'.
    """
    for nameobj, objs in objdict.items():
        prev_scope = None
        for obj, scope in objs:
            if hasattr(obj, name):
                yield quote(nameobj + '.' + name) + \
                    ('' if prev_scope is None else
                     ' ({0} hidden by {1})'.format(scope, prev_scope))
                break
            prev_scope = scope


def suggest_name_as_missing_import(name, objdict, frame):
    """Suggest that name could come from missing import.

    Example: 'foo' -> 'import mod, mod.foo'.
    """
    for mod in STAND_MODULES:
        if mod not in objdict and name in dir(import_from_frame(mod, frame)):
            yield "'{0}' from {1} (not imported)".format(name, mod)


def suggest_name_as_standard_module(name):
    """Suggest that name could be a non-imported standard module.

    Example: 'os.whatever' -> 'import os' and then 'os.whatever'.
    """
    if name in STAND_MODULES:
        yield 'to import {0} first'.format(name)


def suggest_name_as_name_typo(name, objdict):
    """Suggest that name could be a typo (misspelled existing name).

    Example: 'foobaf' -> 'foobar'.
    """
    for name in get_close_matches(name, objdict.keys()):
        yield quote(name) + ' (' + objdict[name][0].scope + ')'


def suggest_name_as_keyword_typo(name):
    """Suggest that name could be a typo (misspelled keyword).

    Example: 'yieldd' -> 'yield'.
    """
    for name in get_close_matches(name, keyword.kwlist):
        yield quote(name) + " (keyword)"


def suggest_name_as_special_case(name):
    """Suggest that name could correspond to a typo with special handling."""
    special_cases = {
        # Imaginary unit is '1j' in Python
        'i': quote('1j') + " (imaginary unit)",
        'j': quote('1j') + " (imaginary unit)",
        # Shell commands entered in interpreter
        'pwd': quote('os.getcwd()'),
        'ls': quote('os.listdir(os.getcwd())'),
        'cd': quote('os.chdir(path)'),
        'rm': "'os.remove(filename)', 'shutil.rmtree(dir)' for recursive",
    }
    result = special_cases.get(name)
    if result is not None:
        yield result


def example_of_suggest_function(value, frame, groups):
    """Example of function to provide suggestion.

    Most functions below are called in a generic way from:
        def get_suggestions_for_error(value, frame, re_dict):
    via
        def get_suggestions_for_exception(value, traceback):
    and thus need to have the same signature and behavior.
    The function is expected to yield any number (0 included) of suggestions
    (as string).
    The parameters are: (value, frame, groups):
     - value: Exception object
     - frame: Last frame of the traceback (may be None when the traceback is
        None which happens only in edge cases)
     - groups: Groups from the error message matched by the error message.
    """
    yield 'this is a suggestion'
    yield 'this is another suggestion'


# Functions related to AttributeError
def suggest_attribute_error(value, frame, groups):
    """Get suggestions in case of ATTRIBUTEERROR."""
    del value  # unused param
    type_str, attr = groups
    return get_attribute_suggestions(type_str, attr, frame)


def suggest_module_has_no_attr(value, frame, groups):
    """Get suggestions in case of MODULEHASNOATTRIBUTE."""
    del value  # unused param
    _, attr = groups  # name ignored for the time being
    return get_attribute_suggestions('module', attr, frame)


def get_attribute_suggestions(type_str, attribute, frame):
    """Get the suggestions closest to the attribute name for a given type."""
    types = get_types_for_str(type_str, frame)
    attributes = set(a for t in types for a in dir(t))
    if type_str == 'module':
        # For module, we manage to get the corresponding 'module' type
        # but the type doesn't bring much information about its content.
        # A hacky way to do so is to assume that the exception was something
        # like 'module_name.attribute' so that we can actually find the module
        # based on the name. Eventually, we check that the found object is a
        # module indeed. This is not failproof but it brings a whole lot of
        # interesting suggestions and the (minimal) risk is to have invalid
        # suggestions.
        module_name = frame.f_code.co_names[0]
        objs = get_objects_in_frame(frame)
        mod = objs[module_name][0].obj
        attributes = set(dir(mod))

    return itertools.chain(
        suggest_attribute_as_builtin(attribute, type_str, frame),
        suggest_attribute_as_removed(attribute, type_str, attributes),
        suggest_attribute_synonyms(attribute, attributes),
        suggest_attribute_as_typo(attribute, attributes))


def suggest_attribute_as_builtin(attribute, type_str, frame):
    """Suggest that a builtin was used as an attribute.

    Example: 'lst.len()' -> 'len(lst)'.
    """
    if attribute in frame.f_builtins:
        yield quote(attribute + '(' + type_str + ')')


def suggest_attribute_as_removed(attribute, type_str, attributes):
    """Suggest that attribute is removed (and give an alternative)."""
    if attribute == 'has_key' and '__contains__' in attributes:
        yield quote('key in ' + type_str)


def suggest_attribute_synonyms(attribute, attributes):
    """Suggest that a method with a similar meaning was used.

    Example: 'lst.add(e)' -> 'lst.append(e)'.
    """
    for set_sub in SYNONYMS_SETS:
        if attribute in set_sub:
            for syn in set_sub & attributes:
                yield quote(syn)


def suggest_attribute_as_typo(attribute, attributes):
    """Suggest the attribute could be a typo.

    Example: 'a.do_baf()' -> 'a.do_bar()'.
    """
    for name in get_close_matches(attribute, attributes):
        # Handle Private name mangling
        if name.startswith('_') and '__' in name and not name.endswith('__'):
            yield quote(name) + ' (but it is supposed to be private)'
        else:
            yield quote(name)


# Functions related to ImportError
def suggest_no_module(value, frame, groups):
    """Get the suggestions closest to the failing module import.

    Example: 'import maths' -> 'import math'.
    """
    del value, frame  # unused param
    module_str, = groups
    for name in get_close_matches(module_str, STAND_MODULES):
        yield quote(name)


def suggest_cannot_import(value, frame, groups):
    """Get the suggestions closest to the failing import."""
    del value  # unused param
    imported_name, = groups
    module_name = frame.f_code.co_names[0]
    return itertools.chain(
        suggest_imported_name_as_typo(imported_name, module_name, frame),
        suggest_import_from_module(imported_name, frame))


def suggest_imported_name_as_typo(imported_name, module_name, frame):
    """Suggest that imported name could be a typo from actual name in module.

    Example: 'from math import pie' -> 'from math import pi'.
    """
    dir_mod = dir(import_from_frame(module_name, frame))
    for name in get_close_matches(imported_name, dir_mod):
        yield quote(name)


def suggest_import_from_module(imported_name, frame):
    """Suggest than name could be found in a standard module.

    Example: 'from itertools import pi' -> 'from math import pi'.
    """
    for mod in STAND_MODULES:
        if imported_name in dir(import_from_frame(mod, frame)):
            yield quote('from {0} import {1}'.format(mod, imported_name))


# Functions related to TypeError
def suggest_unsubscriptable(value, frame, groups):
    """Get suggestions in case of UNSUBSCRIBTABLE error."""
    del value  # unused param
    type_str, = groups
    types = get_types_for_str(type_str, frame)
    if any(hasattr(t, '__call__') for t in types):
        yield quote(type_str + '(value)')


def suggest_unexpected_keywordarg(value, frame, groups):
    """Get suggestions in case of UNEXPECTED_KEYWORDARG error."""
    del value  # unused param
    func_name, kw_arg = groups
    objs = get_objects_in_frame(frame)
    if func_name in objs:
        func = objs[func_name][0].obj
        args = func.__code__.co_varnames
        for name in get_close_matches(kw_arg, args):
            yield quote(name)


def suggest_unexpected_keywordarg2(value, frame, groups):
    """Get suggestions in case of UNEXPECTED_KEYWORDARG2 error."""
    del value
    del frame  # unused param
    del groups  # unused param
    return []  # no implementation so far


def suggest_not_callable(value, frame, groups):
    """Get suggestions in case of NOT_CALLABLE error."""
    del value  # unused param
    type_str, = groups
    types = get_types_for_str(type_str, frame)
    if any(hasattr(t, '__getitem__') for t in types):
        yield quote(type_str + '[value]')


# Functions related to ValueError
def suggest_zero_len_field(value, frame, groups):
    """Get suggestions in case of ZERO_LEN_FIELD."""
    del value, frame, groups  # unused param
    yield '{0}'


def suggest_time_data_is_wrong(value, frame, groups):
    """Get suggestions in case of TIME_DATA_DOES_NOT_MATCH_FORMAT_RE."""
    del value, frame
    timedata, timeformat = groups
    if timedata.count('%') > timeformat.count('%%'):
        yield "to swap value and format parameters"


# Functions related to SyntaxError
def suggest_outside_func_error(value, frame, groups):
    """Get suggestions in case of OUTSIDE_FUNCTION error."""
    del value, frame  # unused param
    yield "to indent it"
    word, = groups
    if word == 'return':
        yield "'sys.exit([arg])'"


def suggest_future_feature(value, frame, groups):
    """Get suggestions in case of FUTURE_FEATURE_NOT_DEF error."""
    del value  # unused param
    feature, = groups
    return suggest_imported_name_as_typo(feature, '__future__', frame)


def suggest_invalid_comp(value, frame, groups):
    """Get suggestions in case of INVALID_COMP error."""
    del value, frame, groups  # unused param
    yield quote('!=')


def suggest_invalid_syntax(value, frame, groups):
    """Get suggestions in case of INVALID_SYNTAX error."""
    del frame, groups  # unused param
    offset = value.offset
    if offset is not None and offset > 2:
        two_last = value.text[offset - 2:offset]
        if two_last == '<>':
            yield quote('!=')


# Functions related to MemoryError
def get_memory_error_sugg(value, frame):
    """Get suggestions for MemoryError exception."""
    assert isinstance(value, MemoryError)
    objs = get_objects_in_frame(frame)
    return itertools.chain.from_iterable(
        suggest_memory_friendly_equi(name, objs)
        for name in frame.f_code.co_names)


# Functions related to OverflowError
def suggest_too_many_items(value, frame, groups):
    """Suggest for TOO_MANY_ITEMS error."""
    del value  # unused param
    func, = groups
    objs = get_objects_in_frame(frame)
    return suggest_memory_friendly_equi(func, objs)


def suggest_memory_friendly_equi(name, objs):
    """Suggest name of a memory friendly equivalent for a function."""
    suggs = {'range': ['xrange']}
    return [quote(s) for s in suggs.get(name, []) if s in objs]


# Functions related to IOError/OSError
def get_io_os_error_sugg(value, frame):
    """Get suggestions for IOError/OSError exception."""
    # https://www.python.org/dev/peps/pep-3151/
    del frame  # unused param
    assert isinstance(value, (IOError, OSError))
    err, _ = value.args
    errnos = {
        errno.ENOENT: suggest_if_file_does_not_exist,
        errno.ENOTDIR: suggest_if_file_is_not_dir,
        errno.EISDIR: suggest_if_file_is_dir,
    }
    return errnos.get(err, lambda x: [])(value)


def suggest_if_file_does_not_exist(value):
    """Get suggestions when a file does not exist."""
    # TODO: Add fuzzy match
    filename = value.filename
    for func, name in (
            (os.path.expanduser, 'os.path.expanduser'),
            (os.path.expandvars, 'os.path.expandvars')):
        expanded = func(filename)
        if os.path.exists(expanded) and filename != expanded:
            yield quote(expanded) + " (calling " + name + ")"


def suggest_if_file_is_not_dir(value):
    """Get suggestions when a file should have been a dir and is not."""
    filename = value.filename
    yield quote(os.path.dirname(filename)) + " (calling os.path.dirname)"


def suggest_if_file_is_dir(value):
    """Get suggestions when a file is a dir and should not."""
    filename = value.filename
    listdir = sorted(os.listdir(filename))
    if listdir:
        trunc_l = listdir[:MAX_NB_FILES]
        truncated = listdir != trunc_l
        filelist = [quote(f) for f in trunc_l] + (["etc"] if truncated else [])
        yield "any of the {0} files in directory ({1})".format(
            len(listdir), ", ".join(filelist))
    else:
        yield "to add content to {0} first".format(filename)


NAMEERRORS = {
    re.VARREFBEFOREASSIGN_RE: suggest_name_not_defined,
    re.NAMENOTDEFINED_RE: suggest_name_not_defined,
}

ATTRIBUTEERRORS = {
    re.ATTRIBUTEERROR_RE: suggest_attribute_error,
    re.MODULEHASNOATTRIBUTE_RE: suggest_module_has_no_attr,
}

TYPEERRORS = {
    re.UNSUBSCRIBTABLE_RE: suggest_unsubscriptable,
    re.UNEXPECTED_KEYWORDARG_RE: suggest_unexpected_keywordarg,
    re.UNEXPECTED_KEYWORDARG2_RE: suggest_unexpected_keywordarg2,
    re.NOT_CALLABLE_RE: suggest_not_callable,
}

VALUEERRORS = {
    re.ZERO_LEN_FIELD_RE: suggest_zero_len_field,
    re.TIME_DATA_DOES_NOT_MATCH_FORMAT_RE: suggest_time_data_is_wrong,
}

SYNTAXERRORS = {
    re.OUTSIDE_FUNCTION_RE: suggest_outside_func_error,
    re.FUTURE_FEATURE_NOT_DEF_RE: suggest_future_feature,
    re.INVALID_COMP_RE: suggest_invalid_comp,
    re.INVALID_SYNTAX_RE: suggest_invalid_syntax,
}

OVERFLOWERRORS = {
    re.RESULT_TOO_MANY_ITEMS_RE: suggest_too_many_items
}

IMPORTERRORS = {
    re.NOMODULE_RE: suggest_no_module,
    re.CANNOTIMPORT_RE: suggest_cannot_import,
}


def get_suggestions_for_error(value, frame, re_dict):
    """Get suggestions for a given error `value`.

    Call relevant suggestion generator based on error message
    and regex dictionnary.
    """
    error_msg = value.args[0]
    for regex, generator in re_dict.items():
        match = re.match(regex, error_msg)
        if match:
            return generator(value, frame, match.groups())
    return []


def get_suggestions_for_exception(value, traceback):
    """Get suggestions for an exception."""
    frame = get_last_frame(traceback)

    def partial_app(re_dict):
        """Partial application of get_suggestions_for_error to re_dict."""
        return (lambda val, fram:
                get_suggestions_for_error(val, fram, re_dict))

    error_types = {
        NameError: partial_app(NAMEERRORS),
        AttributeError: partial_app(ATTRIBUTEERRORS),
        TypeError: partial_app(TYPEERRORS),
        ValueError: partial_app(VALUEERRORS),
        ImportError: partial_app(IMPORTERRORS),
        SyntaxError: partial_app(SYNTAXERRORS),
        OverflowError: partial_app(OVERFLOWERRORS),
        MemoryError: get_memory_error_sugg,
        (IOError, OSError): get_io_os_error_sugg,
    }
    return itertools.chain.from_iterable(
        generator(value, frame)
        for error_type, generator in error_types.items()
        if isinstance(value, error_type))


def add_string_to_exception(value, string):
    """Add string to the exception parameter."""
    # The point is to have the string visible when the exception is printed
    # or converted to string - may it be via `str()`, `repr()` or when the
    # exception is uncaught and displayed (which seems to use `str()`).
    # In an ideal world, one just needs to update `args` but apparently it
    # is not enough for SyntaxError, IOError, etc where other
    # attributes (`msg`, `strerror`, `reason`, etc) are to be updated too
    # (for `str()`, not for `repr()`).
    # Also, elements in args might not be strings or args might me empty
    # so we add to the first string and add the element otherwise.
    assert type(value.args) == tuple
    if string:
        lst_args = list(value.args)
        for i, arg in enumerate(lst_args):
            if isinstance(arg, str):
                lst_args[i] = arg + string
                break
        else:
            # if no string arg, add the string anyway
            lst_args.append(string)
        value.args = tuple(lst_args)
        for attr in ['msg', 'strerror', 'reason']:
            attrval = getattr(value, attr, None)
            if attrval is not None:
                setattr(value, attr, attrval + string)


def get_last_frame(traceback):
    """Extract last frame from a traceback."""
    # In some rare case, the give traceback might be None
    if traceback is None:
        return None
    while traceback.tb_next:
        traceback = traceback.tb_next
    return traceback.tb_frame


def add_suggestions_to_exception(type_, value, traceback):
    """Add suggestion to an exception.

    Arguments are such as provided by sys.exc_info().
    """
    assert isinstance(value, type_)
    add_string_to_exception(
        value,
        get_suggestion_string(
            get_suggestions_for_exception(
                value,
                traceback)))

SylvainDe / DidYouMean-Python

Push — master ( 84ff32...ae1382 )

example_of_suggest_function() A

Complexity

Size

Duplication

Importance

Duplication Side-by-Side

Filter issues like
