def load(self,gramSpec,allResults=0,hypothesis=0, grammarName=None):
if type(gramSpec) == types.StringType:
gramSpec = [gramSpec]
elif type(gramSpec) != types.ListType:
raise TypeError( "grammar definition must be a list of strings" )
elif type(gramSpec[0]) != types.StringType:
raise TypeError( "grammar definition must be a list of strings" )
gramparser.splitApartLines(gramSpec)
parser = gramparser.GramParser(gramSpec, grammarName=grammarName)
parser.doParse()
parser.checkForErrors()
gramBin = gramparser.packGrammar(parser)
self.scanObj = parser.scanObj # for later error messages.
try:
GramClassBase.load(self,gramBin,allResults,hypothesis)
except natlink.BadGrammar:
print 'GrammarBase, cannot load grammar, BadGrammar:\n%s\n'% gramSpec
raise
# we want to keep a list of the rules which can be activated and the
# known lists so we can catch errors earlier
self.validRules = parser.exportRules.keys()
self.validLists = parser.knownLists.keys()
# we reverse the rule dictionary so we can convert rule numbers back
# to rule names during recognition
self.ruleMap = {}
for x in parser.knownRules.keys():
self.ruleMap[ parser.knownRules[x] ] = x
return 1
def load(self, gramSpec, allResults=0, hypothesis=0, grammarName=None):
if type(gramSpec) == types.StringType:
gramSpec = [gramSpec]
elif type(gramSpec) != types.ListType:
raise TypeError("grammar definition must be a list of strings")
elif type(gramSpec[0]) != types.StringType:
raise TypeError("grammar definition must be a list of strings")
gramparser.splitApartLines(gramSpec)
parser = gramparser.GramParser(gramSpec, grammarName=grammarName)
parser.doParse()
parser.checkForErrors()
gramBin = gramparser.packGrammar(parser)
self.scanObj = parser.scanObj # for later error messages.
GramClassBase.load(self, gramBin, allResults, hypothesis)
# we want to keep a list of the rules which can be activated and the
# known lists so we can catch errors earlier
self.validRules = parser.exportRules.keys()
self.validLists = parser.knownLists.keys()
# we reverse the rule dictionary so we can convert rule numbers back
# to rule names during recognition
self.ruleMap = {}
for x in parser.knownRules.keys():
self.ruleMap[parser.knownRules[x]] = x
return 1
def ParseGrammarDefinitions(gramSpec,
gramName,
Lists,
Dicts,
activeRules,
All=1,
Exclusive=0,
exclusiveState=0):
if type(gramSpec) != type([]): gramSpec = [gramSpec]
gramparser.splitApartLines(gramSpec)
## Parser = natlinkutils.GramParser(gramSpec)
## Parser.doParse()
# with gramparserlexyacc:
Parser = gramparser.GramParser(gramSpec)
print '%s, type gramSpec: %s, ' % (gramName, type(gramSpec))
if type(gramSpec) == types.ListType:
gramSpec = [checkForBinary(g) for g in gramSpec]
Parser.doParse()
ParserInfo = (InverseDict(Parser.knownWords),
InverseDict(Parser.knownRules),
InverseDict(Parser.knownLists), Parser.importRules)
stack = []
for name in Parser.ruleDefines.keys():
ParseRuleDefinitions(name, stack, Parser, ParserInfo, Lists, Dicts)
DefRules = {}
for x in stack:
DefRules[x.Name] = x
UsedRules = []
for x in stack:
x.FillInRules(DefRules, UsedRules)
x.FoldLongAlternatives(0)
Grammar = GrammarElement()
Grammar.Init(RuleCode, gramName)
if Exclusive:
if not exclusiveState:
return
# if asking for exclusive, only show the activerules
All = 0
if All:
Obsolete = GrammarElement()
Obsolete.Init(RuleCode, 'Obsolete')
for rule in stack:
if Parser.exportRules.has_key(rule.Name):
Grammar.Insert(rule)
elif not rule.Name in UsedRules:
Obsolete.Insert(rule)
if len(Obsolete.Included) != 0: Grammar.Append(Obsolete)
elif activeRules:
for rule in stack:
if (rule.Name in activeRules):
Grammar.Insert(rule)
else:
return # nothing if no active rules QH
return Grammar
def ParseGrammarDefinitions(gramSpec, gramName, Lists, Dicts, activeRules, All=1, Exclusive=0, exclusiveState=0):
if type(gramSpec) != type([]):
gramSpec = [gramSpec]
gramparser.splitApartLines(gramSpec)
## Parser = natlinkutils.GramParser(gramSpec)
## Parser.doParse()
# with gramparserlexyacc:
Parser = gramparser.GramParser(gramSpec)
Parser.doParse()
ParserInfo = (
InverseDict(Parser.knownWords),
InverseDict(Parser.knownRules),
InverseDict(Parser.knownLists),
Parser.importRules,
)
stack = []
for name in Parser.ruleDefines.keys():
ParseRuleDefinitions(name, stack, Parser, ParserInfo, Lists, Dicts)
DefRules = {}
for x in stack:
DefRules[x.Name] = x
UsedRules = []
for x in stack:
x.FillInRules(DefRules, UsedRules)
x.FoldLongAlternatives(0)
Grammar = GrammarElement()
Grammar.Init(RuleCode, gramName)
if Exclusive:
if not exclusiveState:
return
# if asking for exclusive, only show the activerules
All = 0
if All:
Obsolete = GrammarElement()
Obsolete.Init(RuleCode, "Obsolete")
for rule in stack:
if Parser.exportRules.has_key(rule.Name):
Grammar.Insert(rule)
elif not rule.Name in UsedRules:
Obsolete.Insert(rule)
if len(Obsolete.Included) != 0:
Grammar.Append(Obsolete)
elif activeRules:
for rule in stack:
if rule.Name in activeRules:
Grammar.Insert(rule)
else:
return # nothing if no active rules QH
return Grammar
def load(self, gramSpec, allResults=0, hypothesis=0, grammarName=None):
# print 'loading grammar %s, gramspec type: %s'% (grammarName, type(gramSpec))
# code upper ascii characters with latin1 if they were in the process entered as unicode
if not type(gramSpec) in (six.text_type, six.binary_type,
types.ListType):
raise TypeError(
"grammar definition of %s must be a string or a list of strings, not %s"
% (grammarName, type(gramSpec)))
# print 'loading %s, type: %s'% (grammarName, type(gramSpec) )
if type(gramSpec) == types.ListType:
for i, grampart in enumerate(gramSpec):
line = grampart
if type(line) == six.binary_type:
line = utilsqh.convertToUnicode(line)
if type(line) == six.text_type:
line = utilsqh.convertToBinary(line)
if line != grampart:
gramSpec[i] = line
if type(gramSpec) == six.binary_type:
gramSpec = utilsqh.convertToUnicode(gramSpec)
if type(gramSpec) == six.text_type:
gramSpec = utilsqh.convertToBinary(gramSpec)
gramSpec = gramSpec.split('\n')
gramSpec = [g.rstrip() for g in gramSpec]
gramparser.splitApartLines(gramSpec)
parser = gramparser.GramParser(gramSpec, grammarName=grammarName)
parser.doParse()
parser.checkForErrors()
gramBin = gramparser.packGrammar(parser)
self.scanObj = parser.scanObj # for later error messages.
try:
GramClassBase.load(self, gramBin, allResults, hypothesis)
except natlink.BadGrammar:
print 'GrammarBase, cannot load grammar, BadGrammar:\n%s\n' % gramSpec
raise
# we want to keep a list of the rules which can be activated and the
# known lists so we can catch errors earlier
self.validRules = parser.exportRules.keys()
self.validLists = parser.knownLists.keys()
# we reverse the rule dictionary so we can convert rule numbers back
# to rule names during recognition
self.ruleMap = {}
for x in parser.knownRules.keys():
self.ruleMap[parser.knownRules[x]] = x
return 1
def load(self, gramSpec, allResults=0, hypothesis=0, grammarName=None):
# print 'loading grammar %s, gramspec type: %s'% (grammarName, type(gramSpec))
# code upper ascii characters with latin1 if they were in the process entered as unicode
if type(gramSpec) == types.ListType:
for grampart in gramSpec:
if type(grampart) == types.UnicodeType:
newGramSpec = [g.encode('latin1') for g in gramSpec]
gramSpec = newGramSpec
break
elif type(gramSpec) == types.UnicodeType:
gramSpec = [gramSpec.encode('latin1')]
elif type(gramSpec) == types.StringType:
gramSpec = [gramSpec]
else:
raise TypeError(
"grammar definition must be a string or a list of strings, not %s"
% gramSpec)
gramparser.splitApartLines(gramSpec)
parser = gramparser.GramParser(gramSpec, grammarName=grammarName)
parser.doParse()
parser.checkForErrors()
gramBin = gramparser.packGrammar(parser)
self.scanObj = parser.scanObj # for later error messages.
try:
GramClassBase.load(self, gramBin, allResults, hypothesis)
except natlink.BadGrammar:
print 'GrammarBase, cannot load grammar, BadGrammar:\n%s\n' % gramSpec
raise
# we want to keep a list of the rules which can be activated and the
# known lists so we can catch errors earlier
self.validRules = parser.exportRules.keys()
self.validLists = parser.knownLists.keys()
# we reverse the rule dictionary so we can convert rule numbers back
# to rule names during recognition
self.ruleMap = {}
for x in parser.knownRules.keys():
self.ruleMap[parser.knownRules[x]] = x
return 1
