def traverseInOrder(item):
result = TypeCheckList(Tree)
result.append(traverseInOrder(item.childlist[0]))
result.append(item)
for childindex in range(1, len(item.childlist)):
result += traverseInOrder(item.childlist[childindex])
return result
def traversePreOrder(item):
from pydsl.Abstract import TypeCheckList
result = TypeCheckList(Tree)
result.append(item)
for child in item.childlist:
result += traversePreOrder(child)
return result
def split(self):
"""splits a result"""
result = TypeCheckList(ParseTree)
for symbol in self.symbollist:
currentlist = self.childlist
while len(currentlist) > 1 and len(currentlist[0].symbollist) > 1 and currentlist[0].symbollist[0] != symbol:
for dpr in currentlist:
if symbol in dpr.symbollist:
currentlist = dpr.childlist
break
break #Not found
assert(len(currentlist < 2))
dpr = currentlist[0]
currentsymbol = dpr.symbollist[0]
if len(dpr.symbollist) > 0:
pass
#Ir eliminando los simbolos de alrededor hasta que se quede solo el que nos interesa junto a su word
if currentsymbol != symbol:
#Not found
raise Exception
result.append(dpr)
return result
def terminal_symbol_reducer(symbol, word, production):
""" Reduces a terminal symbol """
from pydsl.Abstract import TypeCheckList
from pydsl.Grammar.Tree import ParseTree
if not isinstance(word, str):
word = str(word)
validresults = TypeCheckList(ParseTree)
if symbol.boundariesrules.policy == "min":
LOG.debug("terminal_symbol_reducer: policy: min")
for begin in range(0, len(word)):
for end in range(begin, len(word)+1):
if symbol.check(word[begin:end]):
LOG.debug("terminal_symbol_reducer: parsed:"+ str(word[begin:end]))
validresults.append(ParseTree(begin, end, [symbol], word[begin:end], production))
break #found the smallest valid symbol at begin
elif symbol.boundariesrules.policy == "max":
LOG.debug("terminal_symbol_reducer: policy: max")
for begin in range(0, len(word)):
maxword = 0
for end in range(begin, len(word)+1):
if symbol.check(word[begin:end]):
LOG.debug("terminal_symbol_reducer: parsed:"+ str(word[begin:end]))
maxword = end
if maxword > 0:
validresults.append(ParseTree(begin, maxword, [symbol], word[begin:maxword], production))
elif symbol.boundariesrules.policy == "fixed":
LOG.debug("terminal_symbol_reducer: policy: fixed")
size = symbol.boundariesrules.size
for begin in range(0, len(word)):
if symbol.check(word[begin:begin + size]):
LOG.debug("__auxReducer: parsed:"+ str(word[begin:begin + size]))
validresults.append(ParseTree(begin, begin + size, [symbol], word[begin:begin + size], production))
else:
LOG.warning("terminal_symbol_reducer: Unknown size policy")
return TypeCheckList(ParseTree)
return validresults
def __recursive_parser(self, onlysymbol, data, production, showerrors:bool = False, recurssions:int = 0):
import pdb
#pdb.set_trace()
""" Aux function. helps check_word"""
LOG.debug("__recursive_parser: Begin ")
if not isinstance(data, str):
data = str(data)
if recurssions > 16*len(data):
LOG.debug("RECURSSION LIMIT")
return TypeCheckList(ParseTree)
if len(data) == 0:
return TypeCheckList(ParseTree)
#if(len(data) < 4):
# import pdb
# pdb.set_trace()
from ..Symbol import TerminalSymbol, NullSymbol, NonTerminalSymbol
if isinstance(onlysymbol, TerminalSymbol):
#Locate every ocurrence of word and return a set of results. Follow boundariesrules
LOG.debug("Iteration: terminalsymbol")
sproduction = self._productionset.getProductionsBySide([onlysymbol])[0]
result = terminal_symbol_consume(onlysymbol, data, sproduction )
if showerrors and not result:
LOG.debug("error symbolo: " + str(onlysymbol))
#print("ERROR!: " + str(data) + str(onlysymbol))
return TypeCheckList(ParseTree, [ParseTree(0,len(data), [onlysymbol] , data, sproduction, valid = False)])
return result
elif isinstance(onlysymbol, NonTerminalSymbol):
result = TypeCheckList(ParseTree)
tmpresults = []
invalidlist = []
for alternative in self._productionset.getProductionsBySide([onlysymbol]):
alternativetree = RecursiveDescentResultTree(None)
alternativesuccess = True
alternativeinvalidlist = []
for symbolindex, symbol in enumerate(alternative.rightside):
symbolsuccess = False
for totalpos in alternativetree.last_poss():
if totalpos >= len(data):
LOG.debug("Alternative length problem:" + str(alternative) + " Recursion: "+ str(recurssions) + " INTENTO SIMBOLO " + str(symbolindex) + ":" + str(symbol) + " Input: "+ str(data) +" SHIFT " + str(totalpos))
continue
LOG.debug("Alternative:" + str(alternative) + " Recursion: "+ str(recurssions) + " INTENTO SIMBOLO " + str(symbolindex) + ":" + str(symbol) + " Input: "+ str(data) +" SHIFT " + str(totalpos))
if symbol == onlysymbol:
#recurssions += 1
pass
thisresult = self.__recursive_parser(symbol, data[totalpos:], alternative, showerrors, recurssions+1)
#pdb.set_trace()
#print("RECIBE: " + str([(str(x.leftpos),str(x.rightpos)) for x in thisresult]))
allvalids = all([x.valid for x in thisresult])
if thisresult and allvalids:
LOG.debug("Alternative:" + str(alternative) + " Recursion: "+ str(recurssions) + " Trying Symbol " + str(symbolindex) + ":" + str(symbol) + " Input: "+ str(data) +" SHIFT " + str(totalpos)+ " results: " + str(thisresult) + "leftpos: " + str(thisresult[0].leftpos) + " rightpos: " + str(thisresult[-1].rightpos))
symbolsuccess = True
for x in thisresult:
x.shift(totalpos)
exito = alternativetree.append(x, totalpos)
if not exito:
#TODO: Añadir como error al arbol o a otro sitio
LOG.debug("Discarded symbol :" + str(symbol) + " position:" + str(totalpos))
else:
LOG.debug("Added symbol :" + str(symbol) + " position:" + str(totalpos))
else:
for x in thisresult:
if not x.valid:
alternativeinvalidlist.append(x)
if not symbolsuccess:
LOG.debug("Symbol doesn't work" + str(symbol))
alternativesuccess = False
break #Try another alternative
else:
LOG.debug("Symbol work at recursion" + str(recurssions) + ": " + str(symbol))
if not alternativesuccess:
continue #A symbol doesn't work, this alternative doesn't neither
else:
invalidlist += alternativeinvalidlist
for x in alternativetree.get_lists():
LOG.debug("Recurssion" + str(recurssions) + " Adding...",[y.content for y in x])
tmpresults.append(x)
LOG.debug("La iteracion " + str(recurssions) + "result collection finished:" + str(tmpresults))
for alternative in self._productionset.getProductionsBySide([onlysymbol]):
for results in tmpresults:
allvalids = all([x.valid for x in results])
LOG.debug("result: " + str([x.content for x in results]) + ", alternative:" + str(alternative))
nullcounter = alternative.rightside.count(NullSymbol())
nnullresults = 0
for y in [x.symbollist for x in results]:
nnullresults += y.count(NullSymbol())
#print(len(results),nnullresults,len(alternative.rightside),nullcounter)
if len(results) - nnullresults != len(alternative.rightside) - nullcounter:
LOG.debug("Discarded: Bad result number")
continue
if results[-1].rightpos > len(data):
LOG.debug("Discarded: Bad rightpos")
continue
for x in range(min(len(alternative.rightside), len(results))):
#It is the same rule?
if results[x].content != alternative.rightside[x]:
continue
#print([(x.leftpos, x.rightpos) for x in results])
#print(recurssions, results[0].leftpos, results[-1].rightpos, allvalids, data, alternative)
newresult = ParseTree(0, results[-1].rightpos - results[0].leftpos, [onlysymbol], data[results[0].leftpos:results[-1].rightpos], production)
#print(newresult)
for child in results:
newresult.append_child(child)
if not child.valid:
newresult.valid = False
if newresult.valid:
result.append(newresult)
if showerrors and not result:
erroresult = TypeCheckList(ParseTree, [ParseTree(0,len(data), [onlysymbol] , data, production, valid = False)])
for invalid in invalidlist:
if invalid.production.leftside[0] in production.rightside:
erroresult[0].append_child(invalid)
return erroresult
#LOG.debug("Returns: " + str([(str(x.leftpos),str(x.rightpos)) for x in result]))
return result
elif isinstance(onlysymbol, NullSymbol):
return TypeCheckList(ParseTree,[ParseTree(0, 0, [onlysymbol], "", production)])
else:
raise Exception
def __init__(self, leftpos, rightpos, content, valid = True):
self.leftpos = leftpos
self.rightpos = rightpos
self.childlist = TypeCheckList(Tree)
self.content = content
self.valid = valid
class Tree(metaclass = ABCMeta):
def __init__(self, leftpos, rightpos, content, valid = True):
self.leftpos = leftpos
self.rightpos = rightpos
self.childlist = TypeCheckList(Tree)
self.content = content
self.valid = valid
def __bool__(self):
"""checks if it is a null result"""
return self.valid
def __getitem__(self, key, order = "preorder"): #FIXME: getitem and optional argument??
result = []
mylist = self.getAllByOrder(order)
for element in mylist:
if element.content == key:
result.append(element)
return result
def getItemByOrder(self, key, order):
if order == "preorder":
return traversePreOrder(self)
elif order == "inorder":
return traverseInOrder(self)
elif order == "postorder":
return traversePostOrder(self)
else:
raise KeyError
def getAllByOrder(self, order = "preorder"):
if order == "preorder":
return traversePreOrder(self)
elif order == "inorder":
return traverseInOrder(self)
elif order == "postorder":
return traversePostOrder(self)
else:
raise KeyError
def shift(self, amount):
""" shifts position """
self.leftpos += amount
self.rightpos += amount
def append_child(self, dpr):
"""appends dpr to childlist"""
self.childlist.append(dpr)
def __len__(self):
return self.rightpos - self.leftpos
def coverage(self):
if not(self):
return 0, len(self)
if childlist:
childtotal = 0
childcoverage = 0
for child in childlist:
newcoverage, newtotal = child.coverage()
childcoverage += newcoverage
childtotal += newtotal
assert(childtotal == len(self))
return childcoverage, childtotal
else:
return len(self), len(self)
def first_leaf(self):
"""Devuelve la primera hoja a la izq"""
if self.childlist:
return self.childlist[0].first_leaf()
else:
return self
def traversePostOrder(item):
result = TypeCheckList(Tree)
for child in item.childlist:
result += traversePostOrder(child)
result.append(item)
return result
def mix_results(resultll:list, productionset):
""" Mix n sets of results """
from pydsl.Grammar.Tree import ParseTree
production = None
for resultl in resultll:
assert(isinstance(resultl, TypeCheckList) and resultl.instancetype == ParseTree)
midlist = [] #All blocks combinations are stored here
firstindex = 0
while firstindex < len(resultll) and len(resultll[firstindex]) == 0:
firstindex += 1
if firstindex == len(resultll):
return []
validsets = 1
#Processing head set
firstresultl = resultll[firstindex]
for result in firstresultl:
if not(isinstance(result, ParseTree)):
raise TypeError
if result.leftpos == 0:
midlist.append([result])
elif result.leftpos == None:
raise Exception #FIXME:What's the right thing to do here?
#Processing Tail sets
for resultl in resultll[firstindex + 1:]:
#For each result list
if len(resultl) == 0:
continue
for result in resultl:
#for each result
tmp = []
for middleresult in midlist:
#combinamos todos los elementos con los ya apuntados en la lista intermedia
#Here we mix every result with intermediate results list
lastresult = middleresult[-1]
if lastresult.production != result.production:
pass
if result.rightpos == None:
result.rightpos = lastresult.rightpos
result.leftpos = lastresult.rightpos
if lastresult.rightpos == None or result.leftpos == None:
tmp.append(middleresult + [ParseTree(result.leftpos, result.rightpos, \
result.symbollist, result.content, result.production, \
TypeCheckList(ParseTree, result.childlist), result.valid)])
elif lastresult.rightpos == result.leftpos:
tmp.append(middleresult + [ParseTree(result.leftpos, result.rightpos, \
result.symbollist, result.content, result.production, \
TypeCheckList(ParseTree, result.childlist), result.valid)])
midlist += tmp
validsets += 1
#eliminamos los resultados intermedios que no contienen tantos elementos como hemos procesado. Es decir, no se ha encontrado una combinacion valida en la ultima mezcla
#Removes all results that have less elements than the number of valid sets
for element in midlist[:]:
if len(element) != validsets:
midlist.remove(element)
#Combinamos resultados en la lista final
#We mix all results into final result
finallist = TypeCheckList(ParseTree)
for middleresult in midlist:
if len(middleresult) == 1:
finallist.append(middleresult[0])
elif middleresult[0].leftpos != None and middleresult[-1].rightpos != None:
#mezclamos la coleccion y dejamos los originales como hijos
#Creates a node with all elements, and originals nodes are the childs of the new node
symbollist = []
for element in middleresult:
compoundword += element.content
symbollist += element.symbollist
finalresult = ParseTree(middleresult[0].leftpos, middleresult[-1].rightpos, symbollist, compoundword, middleresult[0].production, valid = result.valid)
psl = middleresult[0].production
#Add childs to result. FIXME El problema es que estamos añadiendo como hijos del nuevo los elementos ya creados
error = False
rightside = []
for child in middleresult:
assert(child != finalresult)
finalresult.append_child(child)
if child.production:
rightside += child.production.leftside
if not child.valid:
finalresult.valid = False #valid status propagates upwards
#if error:
# print([str(x.production) for x in middleresult])
# continue
if productionset:
try:
finalresult.production = productionset.getProductionsBySide(rightside, "right")
except IndexError:
finallist += middleresult #rule not found: we add it unprocessed (non joined version)
else:
finallist.append(finalresult) #rule found; we add binded together version
else:
raise Exception
return finallist
