from lepl.support.node import Node
# The contents of this file are subject to the Mozilla Public License
# (MPL) Version 1.1 (the "License"); you may not use this file except
# in compliance with the License. You may obtain a copy of the License
# at http://www.mozilla.org/MPL/
#
# Software distributed under the License is distributed on an "AS IS"
# basis, WITHOUT WARRANTY OF ANY KIND, either express or implied. See
# the License for the specific language governing rights and
# limitations under the License.
#
# The Original Code is LEPL (http://www.acooke.org/lepl)
# The Initial Developer of the Original Code is Andrew Cooke.
# Portions created by the Initial Developer are Copyright (C) 2009-2010
# Andrew Cooke (andrew@acooke.org). All Rights Reserved.
#
# Alternatively, the contents of this file may be used under the terms
# of the LGPL license (the GNU Lesser General Public License,
# http://www.gnu.org/licenses/lgpl.html), in which case the provisions
# of the LGPL License are applicable instead of those above.
#
# If you wish to allow use of your version of this file only under the
# terms of the LGPL License and not to allow others to use your version
# of this file under the MPL, indicate your decision by deleting the
# provisions above and replace them with the notice and other provisions
# required by the LGPL License. If you do not delete the provisions
# above, a recipient may use your version of this file under either the
# MPL or the LGPL License.
'''
Support for managing sets of intervals (not all used).
An interval is an open range implemented as a tuple pair. For example (2,5)
is an interval that represents the integers 2,3,4 and 5.
'''
from bisect import bisect_left
from collections import deque
class Intervals(object):
'''
A set of possible values for a character, described as a collection of
intervals. Each interval is [a, b] (ie a <= x <= b, where x is a character
code). We use open bounds to avoid having to specify an "out of range"
value, making it easier to work with a variety of alphabets.
The intervals are stored in a list, ordered by a, joining overlapping
intervals as necessary.
'''
# pylint: disable-msg=C0103
# (use (a,b) variables consistently)
def __init__(self, intervals, alphabet):
self.alphabet = alphabet
self.__intervals = deque()
for interval in intervals:
self.__append(interval)
self.__intervals = list(self.__intervals)
self.__str = alphabet.fmt_intervals(self.__intervals)
self.__index = []
self.state = None
self.__build_index()
def append(self, interval):
'''
Add an interval to the range.
'''
self.__append(interval)
self.__build_index()
def __build_index(self):
'''
Pre-construct the index used for bisection.
'''
self.__index = [interval[1] for interval in self.__intervals]
def __append(self, interval):
'''
Add an interval to the existing intervals.
This maintains self.__intervals in the normalized form described above.
'''
(a1, b1) = interval
if b1 < a1:
(a1, b1) = (b1, a1)
intervals = deque()
done = False
while self.__intervals:
# pylint: disable-msg=E1103
# (pylint fails to infer type)
(a0, b0) = self.__intervals.popleft()
if a0 <= a1:
if b0 < a1 and (b0 != self.alphabet.before(a1)):
# old interval starts and ends before new interval
# so keep old interval and continue
intervals.append((a0, b0))
elif b1 <= b0:
# old interval starts before and ends after new interval
# so keep old interval, discard new interval and slurp
intervals.append((a0, b0))
done = True
break
else:
# old interval starts before new, but partially overlaps
# so discard old interval, extend new interval and continue
# (since it may overlap more intervals...)
(a1, b1) = (a0, b1)
else:
if b1 < a0 and (b1 != self.alphabet.before(a0)):
# new interval starts and ends before old, so add both
# and slurp
intervals.append((a1, b1))
intervals.append((a0, b0))
done = True
break
elif b0 <= b1:
# new interval starts before and ends after old interval
# so discard old and continue (since it may overlap...)
pass
else:
# new interval starts before old, but partially overlaps,
# add extended interval and slurp rest
intervals.append((a1, b0))
done = True
break
if not done:
intervals.append((a1, b1))
intervals.extend(self.__intervals) # slurp remaining
self.__intervals = intervals
def __str__(self):
return self.__str
def __repr__(self):
return self.__str
def len(self):
'''
The number of intervals in the range.
'''
return len(self.__intervals)
def __getitem__(self, index):
return self.__intervals[index]
def __iter__(self):
return iter(self.__intervals)
def __contains__(self, c):
'''
Does the value lie within the intervals?
'''
if self.__index:
index = bisect_left(self.__index, c)
if index < len(self.__intervals):
(a, b) = self.__intervals[index]
return a <= c <= b
return False
def __hash__(self):
return hash(self.__str)
def __eq__(self, other):
# pylint: disable-msg=W0212
# (test for same class)
return isinstance(other, Intervals) and self.__str == other.__str
class _Character(Node):
'''
A set of intervals that is part of a DFA/NFA graph. This is separate
from the `Intervals` instance since we need to clone the node, but can
keep the intervals (necessary so that node equality works correctly when
the same character is used more than once).
'''
def __init__(self, intervals):
super(_Character, self).__init__(intervals)
self.__intervals = intervals
def append(self, interval):
'''
Add an interval to the range.
'''
self.__intervals.append(interval)
def __str__(self):
return str(self.__intervals)
def __repr__(self):
return repr(self.__intervals)
def len(self):
'''
The number of intervals in the range.
'''
return len(self.__intervals)
def __getitem__(self, index):
return self.__intervals[index]
def __iter__(self):
return iter(self.__intervals)
def __contains__(self, c):
return c in self.__intervals
def build(self, graph, src, dest):
'''
Insert within an NFA graph (although at this level, it's not clear it's
NFA).
'''
graph.connect(src, dest, self)
def Character(intervals, alphabet):
'''
Allow simple construction of Character instances.
'''
return _Character(Intervals(intervals, alphabet))
#class Fragments(object):
# '''
# Similar to Character, but each additional interval fragments the list
# of intervals, instead of creating a new merged interval. For example,
# if (3,5) is added to (1,4) and (7,8) then the result will be the
# intervals (1,2), (3,4), (5,5) and (7,8) - the interval (3,4) is the
# overlap between (1,4) and (3,5).
#
# Used internally to combine transitions.
# '''
#
# # pylint: disable-msg=C0103
# # (use (a,b) variables consistently)
#
# def __init__(self, alphabet, characters=None):
# self.alphabet = alphabet
# self.__intervals = deque()
# if characters:
# for character in characters:
# self.append(character)
#
# def append(self, character):
# '''
# Add a character to the intervals.
# '''
# assert type(character) is Character
# for interval in character:
# self.__append(interval)
#
# def __append(self, interval):
# '''
# Add an interval to the existing intervals.
# '''
# (a1, b1) = interval
# if b1 < a1:
# (a1, b1) = (b1, a1)
# intervals = deque()
# alphabet = self.alphabet
# done = False
# while self.__intervals:
# (a0, b0) = self.__intervals.popleft()
# if a0 <= a1:
# if b0 < a1:
# # old interval starts and ends before new interval
# # so keep old interval and continue
# intervals.append((a0, b0))
# elif b1 <= b0:
# # old interval starts before or with and ends after or with
# # new interval
# # so we have one, two or three new intervals
# if a0 < a1:
# # first part of old
# intervals.append((a0, alphabet.before(a1)))
# # common to both
# intervals.append((a1, b1))
# if b1 < b0:
# # last part of old
# intervals.append((alphabet.after(b1), b0))
# done = True
# break
# else:
# # old interval starts before new, but partially overlaps
# # so split old and continue
# # (since it may overlap more intervals...)
# if a0 < a1:
# # first part of old
# intervals.append((a0, alphabet.before(a1)))
# # common to both
# intervals.append((a1, b0))
# a1 = alphabet.after(b0)
# else:
# if b1 < a0:
# # new interval starts and ends before old
# intervals.append((a1, b1))
# intervals.append((a0, b0))
# done = True
# break
# elif b0 <= b1:
# # new interval starts before and ends after or with old
# # interval
# # so split and continue if extends (since last part may
# # overlap...)
# # first part of new
# intervals.append((a1, alphabet.before(a0)))
# # overlap
# intervals.append((a0, b0))
# if b1 > b0:
# a1 = alphabet.after(b0)
# else:
# done = True
# break
# else:
# # new interval starts before old, but partially overlaps,
# # split and slurp rest
# # first part of new
# intervals.append((a1, alphabet.before(a0)))
# # overlap
# intervals.append((a0, b1))
# # last part of old
# intervals.append((alphabet.after(b1), b0))
# done = True
# break
# if not done:
# intervals.append((a1, b1))
# intervals.extend(self.__intervals) # slurp remaining
# self.__intervals = intervals
#
# def len(self):
# '''
# The number of intervals contained.
# '''
# return len(self.__intervals)
#
# def __getitem__(self, index):
# return self.__intervals[index]
#
# def __iter__(self):
# return iter(self.__intervals)
class IntervalMap(dict):
'''
Map from intervals to values.
Note - this is for open intervals! This means it will not work as
expected for continuous variables (which will overlap when two intervals
share a single boundary value). In other words, you cannot store
(1,2) and (2,3) together because both contain 2.
'''
# pylint: disable-msg=C0103
# (use (a,b) variables consistently)
def __init__(self):
super(IntervalMap, self).__init__()
self.__intervals = []
# None is used as a flag to indicate that a new index is needed
self.__index = None
def index(self):
'''
Build the internal indices. Called automatically when necessary.
'''
second = lambda x: x[1]
self.__intervals = list(sorted(self.keys(), key=second))
# pylint: disable-msg=W0141
self.__index = list(map(second, self.__intervals))
def __setitem__(self, interval, value):
# these are rather inefficient, but perhaps useful during development
# assert None == self[interval[0]], 'Overlap'
# assert None == self[interval[1]], 'Overlap'
self.__index = None
super(IntervalMap, self).__setitem__(interval, value)
def __getitem__(self, point):
'''
The argument here is a single value, not an interval.
'''
if self.__index is None:
self.index()
if self.__index:
try:
index = bisect_left(self.__index, point)
except TypeError as e:
from lepl.regexp.core import RegexpError
raise RegexpError(
"""Input characters are inconsistent with the given alphabet.
This error is often triggered by using the default configuration with non-text input; disable with matcher.config.no_compile_to_regexp().
Alternatively, configure with a suitable alphabet.""")
if index < len(self.__index):
# keep interval for identity on retrieval, just in case
(a, b) = interval = self.__intervals[index]
if a <= point <= b:
return super(IntervalMap, self).__getitem__(interval)
return None
def __delitem__(self, interval):
self.__index = None
super(IntervalMap, self).__delitem__(interval)
def __contains__(self, interval):
(a, b) = interval
return self[a] is not None or self[b] is not None
class TaggedFragments(object):
'''
Similar to Fragments, but associates a value with each initial interval;
on retrieval returns a list of all values associated with fragment.
'''
# pylint: disable-msg=C0103
# (use (a,b) variables consistently)
def __init__(self, alphabet):
self.alphabet = alphabet
self.__intervals = deque()
def append(self, character, value):
'''
Add a range and tag.
'''
assert type(character) is _Character
for interval in character:
self.__append(interval, [value])
def __append(self, interval, v1):
'''
Add an interval to the existing intervals.
'''
(a1, b1) = interval
if b1 < a1:
(a1, b1) = (b1, a1)
intervals = deque()
alphabet = self.alphabet
done = False
while self.__intervals:
((a0, b0), v0) = self.__intervals.popleft()
if a0 <= a1:
if b0 < a1:
# old interval starts and ends before new interval
# so keep old interval and continue
intervals.append(((a0, b0), v0))
elif b1 <= b0:
# old interval starts before or with and ends after or with
# new interval
# so we have one, two or three new intervals
if a0 < a1:
# first part of old
intervals.append(((a0, alphabet.before(a1)), v0))
# common to both
intervals.append(((a1, b1), v0+v1))
if b1 < b0:
# last part of old
intervals.append(((alphabet.after(b1), b0), v0))
done = True
break
else:
# old interval starts before new, but partially overlaps
# so split old and continue
# (since new may overlap more intervals...)
if a0 < a1:
# first part of old
intervals.append(((a0, alphabet.before(a1)), v0))
# common to both
intervals.append(((a1, b0), v0+v1))
a1 = alphabet.after(b0)
else:
if b1 < a0:
# new interval starts and ends before old
intervals.append(((a1, b1), v1))
intervals.append(((a0, b0), v0))
done = True
break
elif b0 <= b1:
# new interval starts before and ends after or with old
# interval
# so split and continue if extends (since last part may
# overlap...)
# first part of new
intervals.append(((a1, alphabet.before(a0)), v1))
# old
intervals.append(((a0, b0), v0+v1))
if b1 > b0:
a1 = alphabet.after(b0)
else:
done = True
break
else:
# new interval starts before old, but partially overlaps,
# split and slurp rest
# first part of new
intervals.append(((a1, alphabet.before(a0)), v1))
# overlap
intervals.append(((a0, b1), v0+v1))
# last part of old
intervals.append(((alphabet.after(b1), b0), v0))
done = True
break
if not done:
intervals.append(((a1, b1), v1))
intervals.extend(self.__intervals) # slurp remaining
self.__intervals = intervals
def len(self):
'''
The number of intervals contained.
'''
return len(self.__intervals)
def __getitem__(self, index):
return self.__intervals[index]
def __iter__(self):
return iter(self.__intervals)