"""OptionsClass
Classes:
Option - Holds information about an option
OptionsClass - A collection of options
Abstract:
This module is used to manage "options" managed in user editable files.
This is the implementation of the Options.options globally shared options
object for the SpamBayes project, but is also able to be used to manage
other options required by each application.
The Option class holds information about an option - the name of the
option, a nice name (to display), documentation, default value,
possible values (a tuple or a regex pattern), whether multiple values
are allowed, and whether the option should be reset when restoring to
defaults (options like server names should *not* be).
The OptionsClass class provides facility for a collection of Options.
It is expected that manipulation of the options will be carried out
via an instance of this class.
Experimental or deprecated options are prefixed with 'x-', borrowing the
practice from RFC-822 mail. If the user sets an option like:
[Tokenizer]
x-transmogrify: True
and an 'x-transmogrify' or 'transmogrify' option exists, it is set silently
to the value given by the user. If the user sets an option like:
[Tokenizer]
transmogrify: True
and no 'transmogrify' option exists, but an 'x-transmogrify' option does,
the latter is set to the value given by the users and a deprecation message
is printed to standard error.
To Do:
o Stop allowing invalid options in configuration files
o Find a regex expert to come up with *good* patterns for domains,
email addresses, and so forth.
o str(Option) should really call Option.unconvert since this is what
it does. Try putting that in and running all the tests.
o [See also the __issues__ string.]
o Suggestions?
"""
# This module is part of the spambayes project, which is Copyright 2002-3
# The Python Software Foundation and is covered by the Python Software
# Foundation license.
__credits__ = "All the Spambayes folk."
# blame for the new format: Tony Meyer <ta-meyer@ihug.co.nz>
__issues__ = """Things that should be considered further and by
other people:
We are very generous in checking validity when multiple values are
allowed and the check is a regex (rather than a tuple). Any sequence
that does not match the regex may be used to delimit the values.
For example, if the regex was simply r"[\d]*" then these would all
be considered valid:
"123a234" -> 123, 234
"123abced234" -> 123, 234
"123XST234xas" -> 123, 234
"123 234" -> 123, 234
"123~!@$%^&@234!" -> 123, 234
If this is a problem, my recommendation would be to change the
multiple_values_allowed attribute from a boolean to a regex/None
i.e. if multiple is None, then only one value is allowed. Otherwise
multiple is used in a re.split() to separate the input.
"""
import sys
import os
import shutil
from tempfile import TemporaryFile
try:
import cStringIO as StringIO
except ImportError:
import StringIO
import re
import types
import locale
try:
True, False, bool
except NameError:
# Maintain compatibility with Python 2.2
True, False = 1, 0
def bool(val):
return not not val
__all__ = ['OptionsClass',
'HEADER_NAME', 'HEADER_VALUE',
'INTEGER', 'REAL', 'BOOLEAN',
'SERVER', 'PORT', 'EMAIL_ADDRESS',
'PATH', 'VARIABLE_PATH', 'FILE', 'FILE_WITH_PATH',
'IMAP_FOLDER', 'IMAP_ASTRING',
'RESTORE', 'DO_NOT_RESTORE', 'IP_LIST',
]
MultiContainerTypes = (types.TupleType, types.ListType)
class Option(object):
def __init__(self, name, nice_name="", default=None,
help_text="", allowed=None, restore=True):
self.name = name
self.nice_name = nice_name
self.default_value = default
self.explanation_text = help_text
self.allowed_values = allowed
self.restore = restore
self.delimiter = None
# start with default value
self.set(default)
def display_name(self):
'''A name for the option suitable for display to a user.'''
return self.nice_name
def default(self):
'''The default value for the option.'''
return self.default_value
def doc(self):
'''Documentation for the option.'''
return self.explanation_text
def valid_input(self):
'''Valid values for the option.'''
return self.allowed_values
def no_restore(self):
'''Do not restore this option when restoring to defaults.'''
return not self.restore
def set(self, val):
'''Set option to value.'''
self.value = val
def get(self):
'''Get option value.'''
return self.value
def multiple_values_allowed(self):
'''Multiple values are allowed for this option.'''
return type(self.default_value) in MultiContainerTypes
def is_valid(self, value):
'''Check if this is a valid value for this option.'''
if self.allowed_values is None:
return False
if self.multiple_values_allowed():
return self.is_valid_multiple(value)
else:
return self.is_valid_single(value)
def is_valid_multiple(self, value):
'''Return True iff value is a valid value for this option.
Use if multiple values are allowed.'''
if type(value) in MultiContainerTypes:
for val in value:
if not self.is_valid_single(val):
return False
return True
return self.is_valid_single(value)
def is_valid_single(self, value):
'''Return True iff value is a valid value for this option.
Use when multiple values are not allowed.'''
if type(self.allowed_values) == types.TupleType:
if value in self.allowed_values:
return True
else:
return False
else:
# special handling for booleans, thanks to Python 2.2
if self.is_boolean and (value == True or value == False):
return True
if type(value) != type(self.value) and \
type(self.value) not in MultiContainerTypes:
# This is very strict! If the value is meant to be
# a real number and an integer is passed in, it will fail.
# (So pass 1. instead of 1, for example)
return False
if value == "":
# A blank string is always ok.
return True
avals = self._split_values(value)
# in this case, allowed_values must be a regex, and
# _split_values must match once and only once
if len(avals) == 1:
return True
else:
# either no match or too many matches
return False
def _split_values(self, value):
# do the regex mojo here
if not self.allowed_values:
return ('',)
try:
r = re.compile(self.allowed_values)
except:
print >> sys.stderr, self.allowed_values
raise
s = str(value)
i = 0
vals = ()
while True:
m = r.search(s[i:])
if m is None:
break
vals += (m.group(),)
delimiter = s[i:i + m.start()]
if self.delimiter is None and delimiter != "":
self.delimiter = delimiter
i += m.end()
return vals
def as_nice_string(self, section=None):
'''Summarise the option in a user-readable format.'''
if section is None:
strval = ""
else:
strval = "[%s] " % (section)
strval += "%s - \"%s\"\nDefault: %s\nDo not restore: %s\n" \
% (self.name, self.display_name(),
str(self.default()), str(self.no_restore()))
strval += "Valid values: %s\nMultiple values allowed: %s\n" \
% (str(self.valid_input()),
str(self.multiple_values_allowed()))
strval += "\"%s\"\n\n" % (str(self.doc()))
return strval
def write_config(self, file):
'''Output value in configuration file format.'''
file.write(self.name)
file.write(': ')
file.write(self.unconvert())
file.write('\n')
def convert(self, value):
'''Convert value from a string to the appropriate type.'''
svt = type(self.value)
if svt == type(value):
# already the correct type
return value
if type(self.allowed_values) == types.TupleType and \
value in self.allowed_values:
# already correct type
return value
if self.is_boolean():
if str(value) == "True" or value == 1:
return True
elif str(value) == "False" or value == 0:
return False
raise TypeError, self.name + " must be True or False"
if self.multiple_values_allowed():
# This will fall apart if the allowed_value is a tuple,
# but not a homogenous one...
if isinstance(self.allowed_values, types.StringTypes):
vals = list(self._split_values(value))
else:
if isinstance(value, types.TupleType):
vals = list(value)
else:
vals = value.split()
if len(self.default_value) > 0:
to_type = type(self.default_value[0])
else:
to_type = types.StringType
for i in range(0, len(vals)):
vals[i] = self._convert(vals[i], to_type)
return tuple(vals)
else:
return self._convert(value, svt)
raise TypeError, self.name + " has an invalid type."
def _convert(self, value, to_type):
'''Convert an int, float or string to the specified type.'''
if to_type == type(value):
# already the correct type
return value
if to_type == types.IntType:
return locale.atoi(value)
if to_type == types.FloatType:
return locale.atof(value)
if to_type in types.StringTypes:
return str(value)
raise TypeError, "Invalid type."
def unconvert(self):
'''Convert value from the appropriate type to a string.'''
if type(self.value) in types.StringTypes:
# nothing to do
return self.value
if self.is_boolean():
# A wee bit extra for Python 2.2
if self.value == True:
return "True"
else:
return "False"
if type(self.value) == types.TupleType:
if len(self.value) == 0:
return ""
if len(self.value) == 1:
v = self.value[0]
if type(v) == types.FloatType:
return locale.str(self.value[0])
return str(v)
# We need to separate out the items
strval = ""
# We use a character that is invalid as the separator
# so that it will reparse correctly. We could try all
# characters, but we make do with this set of commonly
# used ones - note that the first one that works will
# be used. Perhaps a nicer solution than this would be
# to specifiy a valid delimiter for all options that
# can have multiple values. Note that we have None at
# the end so that this will crash and die if none of
# the separators works <wink>.
if self.delimiter is None:
if type(self.allowed_values) == types.TupleType:
self.delimiter = ' '
else:
v0 = self.value[0]
v1 = self.value[1]
for sep in [' ', ',', ':', ';', '/', '\\', None]:
# we know at this point that len(self.value) is at
# least two, because len==0 and len==1 were dealt
# with as special cases
test_str = str(v0) + sep + str(v1)
test_tuple = self._split_values(test_str)
if test_tuple[0] == str(v0) and \
test_tuple[1] == str(v1) and \
len(test_tuple) == 2:
break
# cache this so we don't always need to do the above
self.delimiter = sep
for v in self.value:
if type(v) == types.FloatType:
v = locale.str(v)
else:
v = str(v)
strval += v + self.delimiter
strval = strval[:-len(self.delimiter)] # trailing seperator
else:
# Otherwise, we just hope str() will do the job
strval = str(self.value)
return strval
def is_boolean(self):
'''Return True iff the option is a boolean value.'''
# This is necessary because of the Python 2.2 True=1, False=0
# cheat. The valid values are returned as 0 and 1, even if
# they are actually False and True - but 0 and 1 are not
# considered valid input (and 0 and 1 don't look as nice)
# So, just for the 2.2 people, we have this helper function
try:
if type(self.allowed_values) == types.TupleType and \
len(self.allowed_values) > 0 and \
type(self.allowed_values[0]) == types.BooleanType:
return True
return False
except AttributeError:
# If the user has Python 2.2 and an option has valid values
# of (0, 1) - i.e. integers, then this function will return
# the wrong value. I don't know what to do about that without
# explicitly stating which options are boolean
if self.allowed_values == (False, True):
return True
return False
class OptionsClass(object):
def __init__(self):
self.verbose = None
self._options = {}
self.conversion_table = {} # set by creator if they need it.
#
# Regular expressions for parsing section headers and options.
# Lifted straight from ConfigParser
#
SECTCRE = re.compile(
r'\[' # [
r'(?P<header>[^]]+)' # very permissive!
r'\]' # ]
)
OPTCRE = re.compile(
r'(?P<option>[^:=\s][^:=]*)' # very permissive!
r'\s*(?P<vi>[:=])\s*' # any number of space/tab,
# followed by separator
# (either : or =), followed
# by any # space/tab
r'(?P<value>.*)$' # everything up to EOL
)
def update_file(self, filename):
'''Update the specified configuration file.'''
sectname = None
optname = None
out = TemporaryFile()
if os.path.exists(filename):
f = file(filename, "r")
else:
# doesn't exist, so create it - all the changed options will
# be added to it
if self.verbose:
print >> sys.stderr, "Creating new configuration file",
print >> sys.stderr, filename
f = file(filename, "w")
f.close()
f = file(filename, "r")
written = []
vi = ": " # default; uses the one from the file where possible
while True:
line = f.readline()
if not line:
break
# comment or blank line?
if line.strip() == '' or line[0] in '#;':
out.write(line)
continue
if line.split(None, 1)[0].lower() == 'rem' and line[0] in "rR":
# no leading whitespace
out.write(line)
continue
# continuation line?
if line[0].isspace() and sectname is not None and optname:
continue
# a section header or option header?
else:
# is it a section header?
mo = self.SECTCRE.match(line)
if mo:
# Add any missing from the previous section
if sectname is not None:
self._add_missing(out, written, sectname, vi, False)
sectname = mo.group('header')
# So sections can't start with a continuation line
optname = None
if sectname in self.sections():
out.write(line)
# an option line?
else:
mo = self.OPTCRE.match(line)
if mo:
optname, vi, optval = mo.group('option', 'vi', 'value')
if vi in ('=', ':') and ';' in optval:
# ';' is a comment delimiter only if it follows
# a spacing character
pos = optval.find(';')
if pos != -1 and optval[pos-1].isspace():
optval = optval[:pos]
optval = optval.strip()
# allow empty values
if optval == '""':
optval = ''
optname = optname.rstrip().lower()
if self._options.has_key((sectname, optname)):
out.write(optname)
out.write(vi)
newval = self.unconvert(sectname, optname)
out.write(newval.replace("\n", "\n\t"))
out.write('\n')
written.append((sectname, optname))
for sect in self.sections():
self._add_missing(out, written, sect, vi)
f.close()
out.flush()
if self.verbose:
# save a backup of the old file
shutil.copyfile(filename, filename + ".bak")
# copy the new file across
f = file(filename, "w")
out.seek(0)
shutil.copyfileobj(out, f)
out.close()
f.close()
def _add_missing(self, out, written, sect, vi, label=True):
# add any missing ones, where the value does not equal the default
for opt in self.options_in_section(sect):
if not (sect, opt) in written and \
self.get(sect, opt) != self.default(sect, opt):
if label:
out.write('[')
out.write(sect)
out.write("]\n")
label = False
out.write(opt)
out.write(vi)
newval = self.unconvert(sect, opt)
out.write(newval.replace("\n", "\n\t"))
out.write('\n')
written.append((sect, opt))
def load_defaults(self, defaults):
'''Load default values (stored in this module).'''
for section, opts in defaults.items():
for opt in opts:
# If first item of the tuple is a sub-class of Option, then
# instantiate that (with the rest as args). Otherwise,
# assume standard Options class.
klass = Option
args = opt
try:
if issubclass(opt[0], Option):
klass = opt[0]
args = opt[1:]
except TypeError: # opt[0] not a class
pass
o = klass(*args)
self._options[section, o.name] = o
def merge_files(self, file_list):
for f in file_list:
self.merge_file(f)
def convert_and_set(self, section, option, value):
value = self.convert(section, option, value)
self.set(section, option, value)
def merge_file(self, filename):
import ConfigParser
c = ConfigParser.ConfigParser()
c.read(filename)
for sect in c.sections():
for opt in c.options(sect):
value = c.get(sect, opt)
section = sect
option = opt
if not self._options.has_key((section, option)):
if option.startswith('x-'):
# try setting option without the x- prefix
option = option[2:]
if self._options.has_key((section, option)):
self.convert_and_set(section, option, value)
# not an error if an X- option is missing
else:
option = 'x-' + option
# going the other way, if the option has been
# deprecated, set its x-prefixed version and
# emit a warning
if self._options.has_key((section, option)):
self.convert_and_set(section, option, value)
self._report_deprecated_error(section, opt)
else:
print >> sys.stderr, (
"warning: Invalid option %s in"
" section %s in file %s" %
(opt, sect, filename))
else:
self.convert_and_set(section, option, value)
# not strictly necessary, but convenient shortcuts to self._options
def display_name(self, sect, opt):
'''A name for the option suitable for display to a user.'''
return self._options[sect, opt.lower()].display_name()
def default(self, sect, opt):
'''The default value for the option.'''
return self._options[sect, opt.lower()].default()
def doc(self, sect, opt):
'''Documentation for the option.'''
return self._options[sect, opt.lower()].doc()
def valid_input(self, sect, opt):
'''Valid values for the option.'''
return self._options[sect, opt.lower()].valid_input()
def no_restore(self, sect, opt):
'''Do not restore this option when restoring to defaults.'''
return self._options[sect, opt.lower()].no_restore()
def is_valid(self, sect, opt, value):
'''Check if this is a valid value for this option.'''
return self._options[sect, opt.lower()].is_valid(value)
def multiple_values_allowed(self, sect, opt):
'''Multiple values are allowed for this option.'''
return self._options[sect, opt.lower()].multiple_values_allowed()
def is_boolean(self, sect, opt):
'''The option is a boolean value. (Support for Python 2.2).'''
return self._options[sect, opt.lower()].is_boolean()
def convert(self, sect, opt, value):
'''Convert value from a string to the appropriate type.'''
return self._options[sect, opt.lower()].convert(value)
def unconvert(self, sect, opt):
'''Convert value from the appropriate type to a string.'''
return self._options[sect, opt.lower()].unconvert()
def get_option(self, sect, opt):
'''Get an option.'''
if self.conversion_table.has_key((sect, opt)):
sect, opt = self.conversion_table[sect, opt]
return self._options[sect, opt.lower()]
def get(self, sect, opt):
'''Get an option value.'''
if self.conversion_table.has_key((sect, opt.lower())):
sect, opt = self.conversion_table[sect, opt.lower()]
return self.get_option(sect, opt.lower()).get()
def __getitem__(self, key):
return self.get(key[0], key[1])
def set(self, sect, opt, val=None):
'''Set an option.'''
if self.conversion_table.has_key((sect, opt.lower())):
sect, opt = self.conversion_table[sect, opt.lower()]
if self.is_valid(sect, opt, val):
self._options[sect, opt.lower()].set(val)
else:
print >> sys.stderr, ("Attempted to set [%s] %s with invalid"
" value %s (%s)" %
(sect, opt.lower(), val, type(val)))
def set_from_cmdline(self, arg, stream=None):
"""Set option from colon-separated sect:opt:val string.
If optional stream arg is not None, error messages will be displayed
on stream, otherwise KeyErrors will be propagated up the call chain.
"""
sect, opt, val = arg.split(':', 2)
opt = opt.lower()
try:
val = self.convert(sect, opt, val)
except (KeyError, TypeError), msg:
if stream is not None:
self._report_option_error(sect, opt, val, stream, msg)
else:
raise
else:
self.set(sect, opt, val)
def _report_deprecated_error(self, sect, opt):
print >> sys.stderr, (
"Warning: option %s in section %s is deprecated" %
(opt, sect))
def _report_option_error(self, sect, opt, val, stream, msg):
import textwrap
if sect in self.sections():
vopts = self.options(True)
vopts = [v.split(']', 1)[1] for v in vopts
if v.startswith('[%s]'%sect)]
if opt not in vopts:
print >> stream, "Invalid option:", opt
print >> stream, "Valid options for", sect, "are:"
vopts = ', '.join(vopts)
vopts = textwrap.wrap(vopts)
for line in vopts:
print >> stream, ' ', line
else:
print >> stream, "Invalid value:", msg
else:
print >> stream, "Invalid section:", sect
print >> stream, "Valid sections are:"
vsects = ', '.join(self.sections())
vsects = textwrap.wrap(vsects)
for line in vsects:
print >> stream, ' ', line
def __setitem__(self, key, value):
self.set(key[0], key[1], value)
def sections(self):
'''Return an alphabetical list of all the sections.'''
all = []
for sect, opt in self._options.keys():
if sect not in all:
all.append(sect)
all.sort()
return all
def options_in_section(self, section):
'''Return an alphabetical list of all the options in this section.'''
all = []
for sect, opt in self._options.keys():
if sect == section:
all.append(opt)
all.sort()
return all
def options(self, prepend_section_name=False):
'''Return an alphabetical list of all the options, optionally
prefixed with [section_name]'''
all = []
for sect, opt in self._options.keys():
if prepend_section_name:
all.append('[' + sect + ']' + opt)
else:
all.append(opt)
all.sort()
return all
def display(self):
'''Display options in a config file form.'''
output = StringIO.StringIO()
keys = self._options.keys()
keys.sort()
currentSection = None
for sect, opt in keys:
if sect != currentSection:
if currentSection is not None:
output.write('\n')
output.write('[')
output.write(sect)
output.write("]\n")
currentSection = sect
self._options[sect, opt].write_config(output)
return output.getvalue()
def display_full(self, section=None, option=None):
'''Display options including all information.'''
# Given that the Options class is no longer as nice looking
# as it once was, this returns all the information, i.e.
# the doc, default values, and so on
output = StringIO.StringIO()
# when section and option are both specified, this
# is nothing more than a call to as_nice_string
if section is not None and option is not None:
output.write(self._options[section,
option.lower()].as_nice_string(section))
return output.getvalue()
all = self._options.keys()
all.sort()
for sect, opt in all:
if section is not None and sect != section:
continue
output.write(self._options[sect, opt.lower()].as_nice_string(sect))
return output.getvalue()
# These are handy references to commonly used regex/tuples defining
# permitted values. Although the majority of options use one of these,
# you may use any regex or tuple you wish.
HEADER_NAME = r"[\w\.\-\*]+"
HEADER_VALUE = r".+"
INTEGER = r"[\d]+" # actually, a *positive* integer
REAL = r"[\d]+[\.]?[\d]*" # likewise, a *positive* real
BOOLEAN = (False, True)
SERVER = r"([\w\.\-]+(:[\d]+)?)" # in the form server:port
PORT = r"[\d]+"
EMAIL_ADDRESS = r"[\w\-\.]+@[\w\-\.]+"
PATH = r"[\w \$\.\-~:\\/\*\@\=]+"
VARIABLE_PATH = PATH + r"%"
FILE = r"[\S]+"
FILE_WITH_PATH = PATH
IP_LIST = r"\*|localhost|((\*|[01]?\d\d?|2[04]\d|25[0-5])\.(\*|[01]?\d" \
r"\d?|2[04]\d|25[0-5])\.(\*|[01]?\d\d?|2[04]\d|25[0-5])\.(\*" \
r"|[01]?\d\d?|2[04]\d|25[0-5]),?)+"
# IMAP seems to allow any character at all in a folder name,
# but we want to use the comma as a delimiter for lists, so
# we don't allow this. If anyone has folders with commas in the
# names, please let us know and we'll figure out something else.
# ImapUI.py prints out a warning if this is the case.
IMAP_FOLDER = r"[^,]+"
# IMAP's astring should also be valid in the form:
# "{" number "}" CRLF *CHAR8
# where number represents the number of CHAR8 octets
# but this is too complex for us at the moment.
IMAP_ASTRING = ""
for i in range(1, 128):
if not chr(i) in ['"', '\\', '\n', '\r']:
IMAP_ASTRING += chr(i)
IMAP_ASTRING = r"\"?\\?[" + re.escape(IMAP_ASTRING) + r"]+\"?"
# Similarly, each option must specify whether it should be reset to
# this value on a "reset to defaults" command. Most should, but with some
# like a server name that defaults to "", this would be pointless.
# Again, for ease of reading, we define these here:
RESTORE = True
DO_NOT_RESTORE = False
syntax highlighted by Code2HTML, v. 0.9.1