confluent/confluent_server/confluent/noderange.py

# vim: tabstop=4 shiftwidth=4 softtabstop=4

# Copyright 2014 IBM Corporation
# Copyright 2015 Lenovo
#
# Licensed under the Apache License, Version 2.0 (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.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
# this will implement noderange grammar

# considered ast, but a number of things violate python grammar like [] in
# the middle of strings and use of @ for anything is not in their syntax


import itertools
import pyparsing as pp

# construct custom grammar with pyparsing
_nodeword = pp.Word(pp.alphanums + '~^$/=-:.*+!')
_nodebracket = pp.QuotedString(quoteChar='[', endQuoteChar=']',
                               unquoteResults=False)
_nodeatom = pp.Group(pp.OneOrMore(_nodeword | _nodebracket))
_paginationstart = pp.Word('<', pp.nums)
_paginationend = pp.Word('>', pp.nums)
_grammar = _nodeatom | ',-' | ',' | '@' | _paginationstart | _paginationend
_parser = pp.nestedExpr(content=_grammar)

_numextractor = pp.OneOrMore(pp.Word(pp.alphas + '-') | pp.Word(pp.nums))


# TODO: pagination operators <pp.nums and >pp.nums for begin and end respective
class NodeRange(object):
    """Iterate over a noderange

    :param noderange: string representing a noderange to evaluate
    :param config: Config manager object to use to vet elements
    """

    def __init__(self, noderange, config=None):
        self.beginpage = None
        self.endpage = None
        self.cfm = config
        elements = _parser.parseString("(" + noderange + ")").asList()[0]
        self._noderange = self._evaluate(elements)

    @property
    def nodes(self):
        if self.beginpage is None and self.endpage is None:
            return self._noderange
        sortedlist = list(self._noderange)
        sortedlist.sort()
        if self.beginpage is not None:
            sortedlist = sortedlist[self.beginpage:]
        if self.endpage is not None:
            sortedlist = sortedlist[:self.endpage]
        return set(sortedlist)

    def _evaluate(self, parsetree, filternodes=None):
        current_op = 0  # enum, 0 union, 1 subtract, 2 intersect
        current_range = set([])
        if not isinstance(parsetree[0], list):  # down to a plain text thing
            return self._expandstring(parsetree, filternodes)
        for elem in parsetree:
            if elem == ',-':
                current_op = 1
            elif elem == ',':
                current_op = 0
            elif elem == '@':
                current_op = 2
            elif current_op == 0:
                current_range |= self._evaluate(elem)
            elif current_op == 1:
                current_range -= self._evaluate(elem, current_range)
            elif current_op == 2:
                current_range &= self._evaluate(elem, current_range)
        return current_range

    def failorreturn(self, atom):
        if self.cfm is not None:
            raise Exception(atom + " not valid")
        return set([atom])

    def expandrange(self, seqrange, delimiter):
        pieces = seqrange.split(delimiter)
        if len(pieces) % 2 != 0:
            return self.failorreturn(seqrange)
        halflen = len(pieces) / 2
        left = delimiter.join(pieces[:halflen])
        right = delimiter.join(pieces[halflen:])
        leftbits = _numextractor.parseString(left).asList()
        rightbits = _numextractor.parseString(right).asList()
        if len(leftbits) != len(rightbits):
            return self.failorreturn(seqrange)
        finalfmt = ''
        iterators = []
        for idx in xrange(len(leftbits)):
            if leftbits[idx] == rightbits[idx]:
                finalfmt += leftbits[idx]
            elif leftbits[idx][0] in pp.alphas:
                # if string portion unequal, not going to work
                return self.failorreturn(seqrange)
            else:
                curseq = []
                finalfmt += '{%d}' % len(iterators)
                iterators.append(curseq)
                leftnum = int(leftbits[idx])
                rightnum = int(rightbits[idx])
                if leftnum > rightnum:
                    width = len(rightbits[idx])
                    minnum = rightnum
                    maxnum = leftnum + 1  # xrange goes to n-1...
                elif rightnum > leftnum:
                    width = len(leftbits[idx])
                    minnum = leftnum
                    maxnum = rightnum + 1
                else:  # differently padded, but same number...
                    return self.failorreturn(seqrange)
                numformat = '{0:0%d}' % width
                for num in xrange(minnum, maxnum):
                    curseq.append(numformat.format(num))
        results = set([])
        for combo in itertools.product(*iterators):
            entname = finalfmt.format(*combo)
            results |= self.expand_entity(entname)
        return results

    def expand_entity(self, entname):
        if self.cfm is None or self.cfm.is_node(entname):
            return set([entname])
        if self.cfm.is_node(entname):
            return set([entname])
        if self.cfm.is_nodegroup(entname):
            grpcfg = self.cfm.get_nodegroup_attributes(entname)
            nodes = grpcfg['nodes']
            if 'noderange' in grpcfg and grpcfg['noderange']:
                nodes |= NodeRange(
                    grpcfg['noderange']['value'], self.cfm).nodes
            return nodes
        
    def _expandstring(self, element, filternodes=None):
        prefix = ''
        for idx in xrange(len(element)):
            if element[idx][0] == '[':
                nodes = set([])
                for numeric in NodeRange(element[idx][1:-1]).nodes:
                    nodes |= self._expandstring(
                        [prefix + numeric] + element[idx + 1:])
                return nodes
            else:
                prefix += element[idx]
        element = prefix
        if self.cfm is not None:
            # this is where we would check for exactly this
            if self.cfm.is_node(element):
                return set([element])
            if self.cfm.is_nodegroup(element):
                grpcfg = self.cfm.get_nodegroup_attributes(element)
                nodes = grpcfg['nodes']
                if 'noderange' in grpcfg and grpcfg['noderange']:
                    nodes |= NodeRange(
                        grpcfg['noderange']['value'], self.cfm).nodes
                return nodes
        if '-' in element and ':' not in element:
            return self.expandrange(element, '-')
        elif ':' in element:  # : range for less ambiguity
            return self.expandrange(element, ':')
        elif '=' in element or '!~' in element:
            if self.cfm is None:
                raise Exception('Verification configmanager required')
            return set(self.cfm.filter_node_attributes(element, filternodes))
        elif element[0] in ('/', '~'):
            nameexpression = element[1:]
            if self.cfm is None:
                raise Exception('Verification configmanager required')
            return set(self.cfm.filter_nodenames(nameexpression, filternodes))
        elif '+' in element:
            element, increment = element.split('+')
            try:
                nodename, domain = element.split('.')
            except ValueError:
                nodename = element
                domain = ''
            increment = int(increment)
            elembits = _numextractor.parseString(nodename).asList()
            endnum = str(int(elembits[-1]) + increment)
            left = ''.join(elembits)
            if domain:
                left += '.' + domain
            right = ''.join(elembits[:-1])
            right += endnum
            if domain:
                right += '.' + domain
            nrange = left + ':' + right
            return self.expandrange(nrange, ':')
        elif '<' in element:
            self.beginpage = int(element[1:])
            return set([])
        elif '>' in element:
            self.endpage = int(element[1:])
            return set([])
        if self.cfm is None:
            return set([element])
        raise Exception(element + ' not a recognized node, group, or alias')