def run(self, hogwash_job):
print >>sys.stderr, "Loading grammar", self.grammar
grammar = HierGrammar(self.grammar)
print >>sys.stderr, "Done"
print >>sys.stderr, "Parse options:"
print >>sys.stderr, self.parseOpts
self.parseOpts["grammar"] = grammar
if self.parserType == "standard":
parser = Parser(**self.parseOpts)
elif self.parserType == "ctf":
parser = CTFParser(**self.parseOpts)
else:
raise TypeError("Don't know parser type %s" % self.parserType)
print >>sys.stderr, "Parsing:", self.sentence
try:
final = parser.parse(self.sentence)
except (ParseError, TypeError):
#if psyco is active, throwing a parse error will fail
#because psyco doesn't realize that exceptions can be
#newstyle classes, because it's *old*
#so we get a type error
final = parser.parseFail(self.sentence)
print treeToStr(normalizeTree(final.tree()))
return final.tree()
def run(self, hogwash_job):
print >>sys.stderr, "Loading grammar", self.grammar
grammar = HierGrammar(self.grammar)
print >>sys.stderr, "Done"
print >>sys.stderr, "Parse options:"
print >>sys.stderr, self.parseOpts
self.parseOpts["grammar"] = grammar
parser = Parser(**self.parseOpts)
print >>sys.stderr, "Parsing:", self.sentence
try:
final = parser.parse(self.sentence)
res = treeToStr(normalizeTree(final.tree()))
except (ParseError, TypeError):
#if psyco is active, throwing a parse error will fail
#because psyco doesn't realize that exceptions can be
#newstyle classes, because it's *old*
#so we get a type error
final = parser.parseFail(self.sentence)
res = treeToStr(normalizeTree(final.tree()))
print res
return res
def __init__(self, grammar, top="ROOT", queueLimit=10000,
beamF=identityBeamF, mode=None, verbose=[],
makeAnalysis=HierAnalysis,
gammas=[1e-4,], deltas=[1e-3,],
stepExpansionLimit=500,
beamDivergenceFactor=10):
Parser.__init__(self, grammar, top=top, queueLimit=queueLimit,
beamF=beamF, gamma=gammas[0], mode=mode,
verbose=verbose, makeAnalysis=makeAnalysis)
#just for consistency... might have been set with a 'mode' flag
#shouldn't matter though
gammas[0] = self.gamma
self.gammas = gammas
self.deltas = deltas
assert(len(self.deltas) == len(self.gammas) - 1)
self.stepExpansionLimit = stepExpansionLimit
self.beamDivergenceFactor = beamDivergenceFactor
from topdownParser import Grammar, Rule, Parser, normalizeTree, treeToStr
from DBGrammar import DBGrammar
if __name__ == "__main__":
inf = sys.argv[1]
print >>sys.stderr, "loading grammar", inf
grammar = DBGrammar(inf)
print >>sys.stderr, "done"
debug = ["index", "pop", "push", "threshold"]
parser = Parser(grammar, top="ROOT_0", mode="lex",
queueLimit=5e5,
verbose=["index"])
sent = "The stocks fell ."
# sent = "Members of the House Ways and Means Committee introduced legislation that would restrict how the new savings-and-loan bailout agency can raise capital , creating another potential obstacle to the government 's sale of sick thrifts ."
# sent = "The government 's plan"
# sent = "John Smith and Mary Roe are friends ."
#import cProfile
#final = cProfile.run('parser.parse(sent.split())', 'profile-out4')
final = parser.parse(sent.split())
print final
print list(final.derivation())
print treeToStr(final.tree())
print treeToStr(normalizeTree(final.tree()))
from Hogwash import Session #main hogwash class
from Hogwash.Results import ResultsFile #type for file created by hw job
from Hogwash.Action import Action #supertype for runnable objects
from Hogwash.Errors import BadExitCode #error if the program crashed
from waterworks.Processes import bettersystem #run an external command
import sys
from path import path
import os
from shutil import copy
from iterextras import batch
from StringIO import StringIO #store output of process
from distributedParser import Parse
from topdownParser import Parser, Grammar, treeToStr, normalizeTree
if __name__ == "__main__":
session = Session(sys.argv[1], read_only=True, verbose=0)
p = Parser(Grammar({}))
for job in session[:100]:
if job.status != "finished":
sent = job.args[0].sentence
fail = p.parseFail(sent)
print treeToStr(normalizeTree(fail.tree(), unbinarize=True))
else:
print treeToStr(normalizeTree(job.results, unbinarize=False),
epsilonSym="EPSILON")
