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 afterGenerating(self, hyps, i, sentence):
if not hyps[i + 1]:
#fail?
return
for level,gamma in enumerate(self.gammas):
if level == 0:
if "level" in self.verbose:
print >>sys.stderr, "Specified", \
len(hyps[i + 1]), "hypotheses"
bestSub = hyps[i + 1][0]
print >>sys.stderr, "Best parse at level", level,\
treeToStr(bestSub.tree(allowPartial=True)),\
bestSub.prob
print >>sys.stderr, bestSub.expansionProfile(bestSub)
continue
if "level" in self.verbose:
print >>sys.stderr, "Specifying at level", level
self.specifyAtLevel(1, level, hyps[i + 1], 1.0, sentence)
if "level" in self.verbose:
gen = hyps[i + 1]
try:
for subL in range(level):
gen = gen[0].subLevelHeap
bestSub = gen[0]
print >>sys.stderr, "Best parse at level", level,\
treeToStr(bestSub.tree(allowPartial=True)),\
bestSub.prob
print >>sys.stderr, bestSub.expansionProfile(bestSub)
except:
print >>sys.stderr, \
"Warning: best hypothesis has no parse at level",\
level
if self.verbose:
reportMatchingParses(self.verbose, hyps[i + 1], 0)
reportMatchingParses(self.verbose, hyps[i + 1], level)
if "level" in self.verbose:
bestSub = hyps[i + 1][0]
print >>sys.stderr, "Best top-level parse",\
treeToStr(bestSub.tree(allowPartial=True)),\
bestSub.prob
print >>sys.stderr, bestSub
print >>sys.stderr, bestSub.expansionProfile(bestSub)
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()))
print treeToStr(final.tree(True))
# queueLimit=5e5,
verbose=["index", "level", tpar],
gammas=[1e-11, 1e-10, 1e-9, 1e-8],
deltas=[1e-5, 1e-5, 1e-5],
stepExpansionLimit=100,
)
# sent = "The stocks fell ."
# sent = "John Smith and Mary Roe are friends ."
sent = "Food prices are expected to be unchanged , but energy costs jumped as much as 4 % , said Gary Ciminero , economist at Fleet\/Norstar Financial Group ."
final = parser.parse(sent.split())
print final
print list(final.derivation())
print treeToStr(final.tree())
print treeToStr(normalizeTree(final.tree()))
print treeToStr(final.tree(True))
level = 0
while final.subLevelHeap:
level += 1
final = final.subLevelHeap[0]
print
print
print "Level", level
print final
print list(final.derivation())
print treeToStr(final.tree())
print treeToStr(normalizeTree(final.tree(), stripSub=False))
print treeToStr(final.tree(True))
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")