Coverage for ipalib/parameters : 85%
Hot-keys on this page
r m x p toggle line displays
j k next/prev highlighted chunk
0 (zero) top of page
1 (one) first highlighted chunk
|
# Authors: # Jason Gerard DeRose <jderose@redhat.com> # # Copyright (C) 2008 Red Hat # see file 'COPYING' for use and warranty information # # This program is free software; you can redistribute it and/or modify # it under the terms of the GNU General Public License as published by # the Free Software Foundation, either version 3 of the License, or # (at your option) any later version. # # This program is distributed in the hope that it will be useful, # but WITHOUT ANY WARRANTY; without even the implied warranty of # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the # GNU General Public License for more details. # # You should have received a copy of the GNU General Public License # along with this program. If not, see <http://www.gnu.org/licenses/>.
Parameter system for command plugins.
A `Param` instance can be used to describe an argument or option that a command takes, or an attribute that a command returns. The `Param` base class is not used directly, but there are many subclasses for specific Python data types (like `Str` or `Int`) and specific properties (like `Password`).
To create a `Param` instance, you must always provide the parameter *name*, which should be the LDAP attribute name if the parameter describes the attribute of an LDAP entry. For example, we could create an `Str` instance describing the user's last-name attribute like this:
>>> from ipalib import Str >>> sn = Str('sn') >>> sn.name 'sn'
When creating a `Param`, there are also a number of optional kwargs which which can provide additional meta-data and functionality. For example, every parameter has a *cli_name*, the name used on the command-line-interface. By default the *cli_name* is the same as the *name*:
>>> sn.cli_name 'sn'
But often the LDAP attribute name isn't user friendly for the command-line, so you can override this with the *cli_name* kwarg:
>>> sn = Str('sn', cli_name='last') >>> sn.name 'sn' >>> sn.cli_name 'last'
Note that the RPC interfaces (and the internal processing pipeline) always use the parameter *name*, regardless of what the *cli_name* might be.
A `Param` also has two translatable kwargs: *label* and *doc*. These must both be `Gettext` instances. They both default to a place-holder `FixMe` instance, a subclass of `Gettext` used to mark a missing translatable string:
>>> sn.label FixMe('sn') >>> sn.doc FixMe('sn')
The *label* is a short phrase describing the parameter. It's used on the CLI when interactively prompting for values, and as a label for form inputs in the web-UI. The *label* should start with an initial capital. For example:
>>> from ipalib import _ >>> sn = Str('sn', ... cli_name='last', ... label=_('Last name'), ... ) >>> sn.label Gettext('Last name', domain='ipa', localedir=None)
The *doc* is a longer description of the parameter. It's used on the CLI when displaying the help information for a command, and as extra instruction for a form input on the web-UI. By default the *doc* is the same as the *label*:
>>> sn.doc Gettext('Last name', domain='ipa', localedir=None)
But you can override this with the *doc* kwarg. Like the *label*, the *doc* should also start with an initial capital and should not end with any punctuation. For example:
>>> sn = Str('sn', ... cli_name='last', ... label=_('Last name'), ... doc=_("The user's last name"), ... ) >>> sn.doc Gettext("The user's last name", domain='ipa', localedir=None)
Demonstration aside, you should always provide at least the *label* so the various UIs are translatable. Only provide the *doc* if the parameter needs a more detailed description for clarity. """
""" Derive a default value from other supplied values.
For example, say you wanted to create a default for the user's login from the user's first and last names. It could be implemented like this:
>>> login = DefaultFrom(lambda first, last: first[0] + last) >>> login(first='John', last='Doe') 'JDoe'
If you do not explicitly provide keys when you create a `DefaultFrom` instance, the keys are implicitly derived from your callback by inspecting ``callback.func_code.co_varnames``. The keys are available through the ``DefaultFrom.keys`` instance attribute, like this:
>>> login.keys ('first', 'last')
The callback is available through the ``DefaultFrom.callback`` instance attribute, like this:
>>> login.callback # doctest:+ELLIPSIS <function <lambda> at 0x...> >>> login.callback.func_code.co_varnames # The keys ('first', 'last')
The keys can be explicitly provided as optional positional arguments after the callback. For example, this is equivalent to the ``login`` instance above:
>>> login2 = DefaultFrom(lambda a, b: a[0] + b, 'first', 'last') >>> login2.keys ('first', 'last') >>> login2.callback.func_code.co_varnames # Not the keys ('a', 'b') >>> login2(first='John', last='Doe') 'JDoe'
If any keys are missing when calling your `DefaultFrom` instance, your callback is not called and ``None`` is returned. For example:
>>> login(first='John', lastname='Doe') is None True >>> login() is None True
Any additional keys are simply ignored, like this:
>>> login(last='Doe', first='John', middle='Whatever') 'JDoe'
As above, because `DefaultFrom.__call__` takes only pure keyword arguments, they can be supplied in any order.
Of course, the callback need not be a ``lambda`` expression. This third example is equivalent to both the ``login`` and ``login2`` instances above:
>>> def get_login(first, last): ... return first[0] + last ... >>> login3 = DefaultFrom(get_login) >>> login3.keys ('first', 'last') >>> login3.callback.func_code.co_varnames ('first', 'last') >>> login3(first='John', last='Doe') 'JDoe' """
""" :param callback: The callable to call when all keys are present. :param keys: Optional keys used for source values. """ CALLABLE_ERROR % ('callback', callback, type(callback)) ) else: TYPE_ERROR % ('keys', str, key, type(key)) )
self.__class__.__name__, ', '.join(args) )
""" Call the callback if all keys are present.
If all keys are present, the callback is called and its return value is returned. If any keys are missing, ``None`` is returned.
:param kw: The keyword arguments. """
""" Parse shorthand ``spec`` into to ``(name, kw)``.
The ``spec`` string determines the parameter name, whether the parameter is required, and whether the parameter is multivalue according the following syntax:
====== ===== ======== ========== Spec Name Required Multivalue ====== ===== ======== ========== 'var' 'var' True False 'var?' 'var' False False 'var*' 'var' False True 'var+' 'var' True True ====== ===== ======== ==========
For example,
>>> parse_param_spec('login') ('login', {'required': True, 'multivalue': False}) >>> parse_param_spec('gecos?') ('gecos', {'required': False, 'multivalue': False}) >>> parse_param_spec('telephone_numbers*') ('telephone_numbers', {'required': False, 'multivalue': True}) >>> parse_param_spec('group+') ('group', {'required': True, 'multivalue': True})
:param spec: A spec string. """ TYPE_ERROR % ('spec', str, spec, type(spec)) ) '?': dict(required=False, multivalue=False), '*': dict(required=False, multivalue=True), '+': dict(required=True, multivalue=True), }
""" Base class for all parameters.
Param attributes: ================= The behavior of Param class and subclasses can be controlled using the following set of attributes:
- cli_name: option name in CLI - cli_short_name: one character version of cli_name - label: very short description of the parameter. This value is used in when the Command output is printed to CLI or in a Command help - doc: parameter long description used in help - required: the parameter is marked as required for given Command - multivalue: indicates if the attribute is multivalued - primary_key: Command's parameter primary key is used for unique identification of an LDAP object and for sorting - normalizer: a custom function for Param value normalization - default_from: a custom function for generating default values of parameter instance - autofill: by default, only `required` parameters get a default value from the default_from function. When autofill is enabled, optional attributes get the default value filled too - query: this attribute is controlled by framework. When the `query` is enabled, framework assumes that the value is only queried and not inserted in the LDAP. Validation is then relaxed - custom parameter validators are skipped and only basic class validators are executed to check the parameter value - attribute: this attribute is controlled by framework and enabled for all LDAP objects parameters (unless parameter has "virtual_attribute" flag). All parameters with enabled `attribute` are being encoded and placed to an entry passed to LDAP Create/Update calls - include: a list of contexts where this parameter should be included. `Param.use_in_context()` provides further information. - exclude: a list of contexts where this parameter should be excluded. `Param.use_in_context()` provides further information. - flags: there are several flags that can be used to further tune the parameter behavior: * no_display (Output parameters only): do not display the parameter * no_create: do not include the parameter for crud.Create based commands * no_update: do not include the parameter for crud.update based commands * virtual_attribute: the parameter is not stored physically in the LDAP and thus attribute `attribute` is not enabled * suppress_empty (Output parameters only): do not display parameter value when empty * ask_create: CLI asks for parameter value even when the parameter is not `required`. Applied for all crud.Create based commands * ask_update: CLI asks for parameter value even when the parameter is not `required`. Applied for all crud.Update based commands * req_update: The parameter is `required` in all crud.Update based commands * nonempty: This is an internal flag; a required attribute should be used instead of it. The value of this parameter must not be empty, but it may not be given at all. All crud.Update commands automatically convert required parameters to `nonempty` ones, so the value can be unspecified (unchanged) but cannot be deleted. - hint: this attribute is currently not used - alwaysask: when enabled, CLI asks for parameter value even when the parameter is not `required` - sortorder: used to sort a list of parameters for Command. See `Command.finalize()` for further information - csv: this multivalue attribute is given in CSV format - csv_separator: character that separates values in CSV (comma by default) - csv_skipspace: if true, leading whitespace will be ignored in individual CSV values """
# This is a dummy type so that most of the functionality of Param can be # unit tested directly without always creating a subclass; however, a real # (direct) subclass must *always* override this class attribute:
# Subclasses should override this with something more specific:
# _convert_scalar operates only on scalar values
('cli_name', str, None), ('cli_short_name', str, None), ('label', (basestring, Gettext), None), ('doc', (basestring, Gettext), None), ('required', bool, True), ('multivalue', bool, False), ('primary_key', bool, False), ('normalizer', callable, None), ('default_from', DefaultFrom, None), ('autofill', bool, False), ('query', bool, False), ('attribute', bool, False), ('include', frozenset, None), ('exclude', frozenset, None), ('flags', frozenset, frozenset()), ('hint', (str, Gettext), None), ('alwaysask', bool, False), ('sortorder', int, 2), # see finalize() ('csv', bool, False), ('csv_separator', str, ','), ('csv_skipspace', bool, True), ('option_group', unicode, None),
# The 'default' kwarg gets appended in Param.__init__(): # ('default', self.type, None), )
# We keep these values to use in __repr__():
else:
# Merge in kw from parse_param_spec():
# Add 'default' to self.kwargs and makes sure no unknown kw were given: else: '%s: takes no such kwargs: %s' % (self.nice, ', '.join(repr(k) for k in sorted(extra)) ) )
# Merge in default for 'cli_name', label, doc if not given:
# Wrap 'default_from' in a DefaultFrom if not already:
# We keep this copy with merged values also to use when cloning:
# Perform type validation on kw, add in class rules: type(kind) is type and not isinstance(value, kind) or type(kind) is tuple and not isinstance(value, kind) ): TYPE_ERROR % (key, kind, value, type(value)) ) CALLABLE_ERROR % (key, value, type(value)) ) key, self.__class__.__name__) )
# Check that only 'include' or 'exclude' was provided: '%s: cannot have both %s=%r and %s=%r' % ( self.nice, 'include', self.include, 'exclude', self.exclude, ) )
# Check that if csv is set, multivalue is set too
# Check that all the rules are callable # by definition a query enforces no class or parameter rules else: raise TypeError( '%s: rules must be callable; got %r' % (self.nice, rule) )
# Check that cli_short_name is only 1 character long: raise ValueError( '%s: cli_short_name can only be a single character: %s' % ( self.nice, self.cli_short_name) )
# And we're done.
""" Return an expresion that could construct this `Param` instance. """ self.__class__.__name__, ', '.join(self.__repr_iter()) )
else:
""" One stop shopping. """ else: else:
""" Return the right name of an attribute depending on usage.
Normally errors should use cli_name, our "friendly" name. When using the API directly or *attr return the real name. """ name = self.name
""" Iterate through ``(key,value)`` for all kwargs passed to constructor. """ for key in sorted(self.__kw): value = self.__kw[key] if callable(value) and hasattr(value, '__name__'): value = value.__name__ yield (key, value)
""" Return ``True`` if this parameter should be used in ``env.context``.
If a parameter is created with niether the ``include`` nor the ``exclude`` kwarg, this method will always return ``True``. For example:
>>> from ipalib.config import Env >>> param = Param('my_param') >>> param.use_in_context(Env(context='foo')) True >>> param.use_in_context(Env(context='bar')) True
If a parameter is created with an ``include`` kwarg, this method will only return ``True`` if ``env.context`` is in ``include``. For example:
>>> param = Param('my_param', include=['foo', 'whatever']) >>> param.include frozenset(['foo', 'whatever']) >>> param.use_in_context(Env(context='foo')) True >>> param.use_in_context(Env(context='bar')) False
If a paremeter is created with an ``exclude`` kwarg, this method will only return ``True`` if ``env.context`` is not in ``exclude``. For example:
>>> param = Param('my_param', exclude=['foo', 'whatever']) >>> param.exclude frozenset(['foo', 'whatever']) >>> param.use_in_context(Env(context='foo')) False >>> param.use_in_context(Env(context='bar')) True
Note that the ``include`` and ``exclude`` kwargs are mutually exclusive and that at most one can be suppelied to `Param.__init__()`. For example:
>>> param = Param('nope', include=['foo'], exclude=['bar']) Traceback (most recent call last): ... ValueError: Param('nope'): cannot have both include=frozenset(['foo']) and exclude=frozenset(['bar'])
So that subclasses can add additional logic based on other environment variables, the entire `config.Env` instance is passed in rather than just the value of ``env.context``. """
""" Return a value safe for logging.
This is used so that passwords don't get logged. If this is a `Password` instance and ``value`` is not ``None``, a constant ``u'********'`` is returned. For example:
>>> p = Password('my_password') >>> p.safe_value(u'This is my password') u'********' >>> p.safe_value(None) is None True
If this is not a `Password` instance, ``value`` is returned unchanged. For example:
>>> s = Str('my_str') >>> s.safe_value(u'Some arbitrary value') u'Some arbitrary value' """
""" Return a new `Param` instance similar to this one. """
""" Return a new `Param` instance similar to this one, but named differently """
""" Return a new `Param` instance similar to this one, but of a different type """
# The following 2 functions were taken from the Python # documentation at http://docs.python.org/library/csv.html
# csv.py doesn't do Unicode; encode temporarily as UTF-8: dialect=dialect, delimiter=self.csv_separator, quotechar='"', skipinitialspace=self.csv_skipspace, **kwargs) # decode UTF-8 back to Unicode, cell by cell:
"""Split CSV strings into individual values.
For CSV params, ``value`` is a tuple of strings. Each of these is split on commas, and the results are concatenated into one tuple.
For example::
>>> param = Param('telephones', multivalue=True, csv=True) >>> param.split_csv((u'1, 2', u'3', u'4, 5, 6')) (u'1', u'2', u'3', u'4', u'5', u'6')
If ``value`` is not a tuple (or list), it is only split::
>>> param = Param('telephones', multivalue=True, csv=True) >>> param.split_csv(u'1, 2, 3') (u'1', u'2', u'3')
For non-CSV params, return the value unchanged. """ else: newval.append(v) else:
""" Normalize ``value`` using normalizer callback.
For example:
>>> param = Param('telephone', ... normalizer=lambda value: value.replace('.', '-') ... ) >>> param.normalize(u'800.123.4567') u'800-123-4567'
If this `Param` instance was created with a normalizer callback and ``value`` is a unicode instance, the normalizer callback is called and *its* return value is returned.
On the other hand, if this `Param` instance was *not* created with a normalizer callback, if ``value`` is *not* a unicode instance, or if an exception is caught when calling the normalizer callback, ``value`` is returned unchanged.
:param value: A proposed value for this parameter. """ self._normalize_scalar(v) for v in value ) else:
""" Normalize a scalar value.
This method is called once for each value in a multivalue. """
""" Convert ``value`` to the Python type required by this parameter.
For example:
>>> scalar = Str('my_scalar') >>> scalar.type <type 'unicode'> >>> scalar.convert(43.2) u'43.2'
(Note that `Str` is a subclass of `Param`.)
All values in `constants.NULLS` will be converted to ``None``. For example:
>>> scalar.convert(u'') is None # An empty string True >>> scalar.convert([]) is None # An empty list True
Likewise, values in `constants.NULLS` will be filtered out of a multivalue parameter. For example:
>>> multi = Str('my_multi', multivalue=True) >>> multi.convert([1.5, '', 17, None, u'Hello']) (u'1.5', u'17', u'Hello') >>> multi.convert([None, u'']) is None # Filters to an empty list True
Lastly, multivalue parameters will always return a ``tuple`` (assuming they don't return ``None`` as in the last example above). For example:
>>> multi.convert(42) # Called with a scalar value (u'42',) >>> multi.convert([0, 1]) # Called with a list value (u'0', u'1')
Note that how values are converted (and from what types they will be converted) completely depends upon how a subclass implements its `Param._convert_scalar()` method. For example, see `Str._convert_scalar()`.
:param value: A proposed value for this parameter. """ self._convert_scalar(v, i) for (i, v) in filter( lambda iv: iv[1] not in NULLS, enumerate(value) ) )
""" Convert a single scalar value. """ error=ugettext(self.type_error), )
""" Check validity of ``value``.
:param value: A proposed value for this parameter. :param context: The context we are running in. :param supplied: True if this parameter was supplied explicitly. """ else: TYPE_ERROR % ('value', tuple, value, type(value)) ) else:
error='need a %r; got %r (a %r)' % ( self.type, value, type(value) ) ) TYPE_ERROR % ('index', int, index, type(index)) ) name=self.get_param_name(), value=value, index=index, error=error, rule=rule, )
""" Return the static default or construct and return a dynamic default.
(In these examples, we will use the `Str` and `Bytes` classes, which both subclass from `Param`.)
The *default* static default is ``None``. For example:
>>> s = Str('my_str') >>> s.default is None True >>> s.get_default() is None True
However, you can provide your own static default via the ``default`` keyword argument when you create your `Param` instance. For example:
>>> s = Str('my_str', default=u'My Static Default') >>> s.default u'My Static Default' >>> s.get_default() u'My Static Default'
If you need to generate a dynamic default from other supplied parameter values, provide a callback via the ``default_from`` keyword argument. This callback will be automatically wrapped in a `DefaultFrom` instance if it isn't one already (see the `DefaultFrom` class for all the gory details). For example:
>>> login = Str('login', default=u'my-static-login-default', ... default_from=lambda first, last: (first[0] + last).lower(), ... ) >>> isinstance(login.default_from, DefaultFrom) True >>> login.default_from.keys ('first', 'last')
Then when all the keys needed by the `DefaultFrom` instance are present, the dynamic default is constructed and returned. For example:
>>> kw = dict(last=u'Doe', first=u'John') >>> login.get_default(**kw) u'jdoe'
Or if any keys are missing, your *static* default is returned. For example:
>>> kw = dict(first=u'John', department=u'Engineering') >>> login.get_default(**kw) u'my-static-login-default' """ except StandardError: pass
json_dict = {} for (a, k, d) in self.kwargs: if k in (callable, DefaultFrom): continue elif isinstance(getattr(self, a), frozenset): json_dict[a] = [k for k in getattr(self, a, [])] else: json_dict[a] = getattr(self, a, '') json_dict['class'] = self.__class__.__name__ json_dict['name'] = self.name json_dict['type'] = self.type.__name__ return json_dict
""" A parameter for boolean values (stored in the ``bool`` type). """
# FIXME: This my quick hack to get some UI stuff working, change these defaults # --jderose 2009-08-28 ('truths', frozenset, frozenset([1, u'1', True, u'true', u'TRUE'])), ('falsehoods', frozenset, frozenset([0, u'0', False, u'false', u'FALSE'])), )
""" Convert a single scalar value. """ if type(value) in (tuple, list): raise ConversionError(name=self.name, index=index, error=ugettext(self.scalar_error)) raise ConversionError(name=self.name, index=index, error=ugettext(self.type_error), )
""" A boolean parameter that always gets filled in with a default value.
This `Bool` subclass forces ``autofill=True`` in `Flag.__init__()`. If no default is provided, it also fills in a default value of ``False``. Lastly, unlike the `Bool` class, the default must be either ``True`` or ``False`` and cannot be ``None``.
For example:
>>> flag = Flag('my_flag') >>> (flag.autofill, flag.default) (True, False)
To have a default value of ``True``, create your `Flag` intance with ``default=True``. For example:
>>> flag = Flag('my_flag', default=True) >>> (flag.autofill, flag.default) (True, True)
Also note that creating a `Flag` instance with ``autofill=False`` will have no effect. For example:
>>> flag = Flag('my_flag', autofill=False) >>> flag.autofill True """
TYPE_ERROR % ('default', bool, default, type(default)) )
""" Base class for the `Int` and `Decimal` parameters. """
""" Convert a single scalar value. """ if type(value) is self.type: return value if type(value) in (unicode, int, float): try: return self.type(value) except ValueError: pass if type(value) in (tuple, list): raise ConversionError(name=self.name, index=index, error=ugettext(self.scalar_error)) raise ConversionError(name=self.name, index=index, error=ugettext(self.type_error), )
""" A parameter for integer values (stored in the ``int`` type). """
('minvalue', int, int(MININT)), ('maxvalue', int, int(MAXINT)), )
#pylint: disable=E1003
'%s: minvalue > maxvalue (minvalue=%r, maxvalue=%r)' % ( self.nice, self.minvalue, self.maxvalue) )
""" Convert a single scalar value. """ # permit floating point strings else: # 2nd arg is radix base, 2nd arg only accepted for strings. # Zero means determine radix base from prefix (e.g. 0x for hex) except ValueError: pass error=ugettext(self.type_error), )
""" Check min constraint. """ minvalue=self.minvalue, )
""" Check max constraint. """ maxvalue=self.maxvalue, )
""" This duplicates _validate_scalar in the Param class with the exception that it allows both int and long types. The min/max rules handle size enforcement. """ raise ValidationError(name=self.name, error='need a %r; got %r (a %r)' % ( self.type, value, type(value) ) ) raise TypeError( TYPE_ERROR % ('index', int, index, type(index)) ) name=self.get_param_name(), value=value, index=index, error=error, rule=rule, )
""" A parameter for floating-point values (stored in the ``Decimal`` type).
Python Decimal type helps overcome problems tied to plain "float" type, e.g. problem with representation or value comparison. In order to safely transfer the value over RPC libraries, it is being converted to string which is then converted back to Decimal number. """
('minvalue', decimal.Decimal, None), ('maxvalue', decimal.Decimal, None), ('precision', int, None), )
except Exception, e: raise ValueError( '%s: cannot parse kwarg %s: %s' % ( name, kwparam, str(e)))
and (self.minvalue is not None and \ self.maxvalue is not None): '%s: minvalue > maxvalue (minvalue=%s, maxvalue=%s)' % ( self.nice, self.minvalue, self.maxvalue) )
raise ValueError('%s: precision must be at least 0' % self.nice)
""" Check min constraint. """ minvalue=self.minvalue, )
""" Check max constraint. """ maxvalue=self.maxvalue, )
return value
except Exception, e: raise ConversionError(name=self.get_param_name(), index=index, error=unicode(e))
return super(Decimal, self)._convert_scalar(value, index)
""" Base class for the `Bytes` and `Str` parameters.
Previously `Str` was as subclass of `Bytes`. Now the common functionality has been split into this base class so that ``isinstance(foo, Bytes)`` wont be ``True`` when ``foo`` is actually an `Str` instance (which is confusing). """
('minlength', int, None), ('maxlength', int, None), ('length', int, None), ('pattern', (basestring,), None), ('pattern_errmsg', (basestring,), None), )
self.length is None or (self.minlength is None and self.maxlength is None) ): '%s: cannot mix length with minlength or maxlength' % self.nice )
'%s: minlength must be >= 1; got %r' % (self.nice, self.minlength) )
'%s: maxlength must be >= 1; got %r' % (self.nice, self.maxlength) )
'%s: minlength > maxlength (minlength=%r, maxlength=%r)' % ( self.nice, self.minlength, self.maxlength) ) '%s: minlength == maxlength; use length=%d instead' % ( self.nice, self.minlength) )
""" Check pattern (regex) contraint. """ else: pattern=self.pattern, )
""" A parameter for binary data (stored in the ``str`` type).
This class is named *Bytes* instead of *Str* so it's aligned with the Python v3 ``(str, unicode) => (bytes, str)`` clean-up. See:
http://docs.python.org/3.0/whatsnew/3.0.html
Also see the `Str` parameter. """
else:
""" Check minlength constraint. """ minlength=self.minlength, )
""" Check maxlength constraint. """ maxlength=self.maxlength, )
""" Check length constraint. """ length=self.length, )
try: value = base64.b64decode(value) except TypeError: raise ConversionError(name=self.get_param_name(), index=index, error=self.type_error)
""" A parameter for Unicode text (stored in the ``unicode`` type).
This class is named *Str* instead of *Unicode* so it's aligned with the Python v3 ``(str, unicode) => (bytes, str)`` clean-up. See:
http://docs.python.org/3.0/whatsnew/3.0.html
Also see the `Bytes` parameter. """
('noextrawhitespace', bool, True), )
else:
""" Convert a single scalar value. """ error=ugettext(self.scalar_error)) error=ugettext(self.type_error), )
""" Do not allow leading/trailing spaces. """
""" Check minlength constraint. """ minlength=self.minlength, )
""" Check maxlength constraint. """ maxlength=self.maxlength, )
""" Check length constraint. """ length=self.length, )
""" An IA5String per RFC 4517 """
index=index, error=_('The character \'%(char)r\' is not allowed.') % dict(char=value[i],) )
""" A parameter for passwords (stored in the ``unicode`` type). """
('confirm', bool, True), )
""" Base class for parameters with enumerable values. """
('values', tuple, tuple()), )
TYPE_ERROR % (n, self.type, v, type(v)) )
values=self.values, )
""" Enumerable for binary data (stored in the ``str`` type). """
""" Enumerable for Unicode text (stored in the ``unicode`` type).
For example:
>>> enum = StrEnum('my_enum', values=(u'One', u'Two', u'Three')) >>> enum.validate(u'Two', 'cli') is None True >>> enum.validate(u'Four', 'cli') Traceback (most recent call last): ... ValidationError: invalid 'my_enum': must be one of (u'One', u'Two', u'Three') """
""" A parameter capable of holding values of any type. For internal use only. """
error = rule(ugettext, value) if error is not None: raise ValidationError( name=self.name, value=value, index=index, error=error, rule=rule, )
""" File parameter type.
Accepts file names and loads their content into the parameter value. """ # valid for CLI, other backends (e.g. webUI) can ignore this ('stdin_if_missing', bool, False), ('noextrawhitespace', bool, False), )
""" Access time parameter type.
Accepts values conforming to generalizedTime as defined in RFC 4517 section 3.3.13 without time zone information. """ raise ValueError('HHMM must be exactly 4 characters long') raise ValueError('HHMM non-numeric') raise ValueError('HH out of range') raise ValueError('MM out of range')
raise ValueError('invalid day of the week')
raise ValueError('day of the month out of range') elif month_num in (4, 6, 9, 11): if value < 1 or value > 30: raise ValueError('day of the month out of range') elif month_num == 2: if year % 4 == 0 and (year % 100 != 0 or year % 400 == 0): if value < 1 or value > 29: raise ValueError('day of the month out of range') else: if value < 1 or value > 28: raise ValueError('day of the month out of range')
raise ValueError('week of the month non-numeric') raise ValueError('week of the month out of range')
if not t.isnumeric(): raise ValueError('week of the year non-numeric') value = int(t) if value < 1 or value > 52: raise ValueError('week of the year out of range')
if not t.isnumeric(): raise ValueError('day of the year non-numeric') value = int(t) if value < 1 or value > 365: raise ValueError('day of the year out of range')
raise ValueError('month number non-numeric') raise ValueError('month number out of range')
raise ValueError('invalid time range') raise ValueError('invalid time range') raise ValueError('invalid time range')
raise ValueError('invalid week specifier')
else: raise ValueError('invalid month specifier')
elif ts[index] == 'week': self._check_interval(ts[index + 1], self._check_woty) index = self._check_W_spec(ts, index + 2) elif ts[index] == 'day': index += 1 self._check_interval(ts[index], self._check_doty) else: raise ValueError('invalid year specifier')
raise ValueError('incomplete generalized time') raise ValueError('time non-numeric') # don't check year value, with time travel and all :) else: self._check_HHMM('%s00' % t[8:10]) s = int(t[12:14]) if s < 0 or s > 60: raise ValueError('seconds out of range')
raise ValueError('invalid time range') raise ValueError('period must be yearly, monthy or daily, got \'%s\'' % ts[1]) else:
except IndexError: raise ValidationError( name=self.get_param_name(), error='incomplete time value' )
""" Create an `Str` instance from the shorthand ``spec``.
This function allows you to create `Str` parameters (the most common) from a convenient shorthand that defines the parameter name, whether it is required, and whether it is multivalue. (For the definition of the shorthand syntax, see the `parse_param_spec()` function.)
If ``spec`` is an ``str`` instance, it will be used to create a new `Str` parameter, which will be returned. For example:
>>> s = create_param('hometown?') >>> s Str('hometown?') >>> (s.name, s.required, s.multivalue) ('hometown', False, False)
On the other hand, if ``spec`` is already a `Param` instance, it is returned unchanged. For example:
>>> b = Bytes('cert') >>> create_param(b) is b True
As a plugin author, you will not call this function directly (which would be no more convenient than simply creating the `Str` instance). Instead, `frontend.Command` will call it for you when it evaluates the ``takes_args`` and ``takes_options`` attributes, and `frontend.Object` will call it for you when it evaluates the ``takes_params`` attribute.
:param spec: A spec string or a `Param` instance. """ TYPE_ERROR % ('spec', (str, Param), spec, type(spec)) ) |