#this is useless, I think
RuleApplicationHypothesis.set_minimum_covered_words(True)
ruleList=RuleList()
#load alignment templates
if args.alignment_templates:
if args.alignment_templates.endswith(".gz"):
gfile=gzip.open(args.alignment_templates)
else:
gfile=open(args.alignment_templates)
for line in gfile:
line=line.strip().decode('utf-8')
at=AlignmentTemplate()
at.parse(line)
ruleList.add(at)
gfile.close()
#load sentences
sentences=list()
if args.sentences:
if args.sentences.lower().endswith('.gz'):
gfile=gzip.open(args.sentences)
else:
gfile=open(args.sentences)
for line in gfile:
line=line.strip().decode('utf-8')
parallelSentence=ParallelSentence()
parallelSentence.parse(line, parseTlLemmasFromDic=True,forRBPE=args.rbpe)
if not args.tt1_beam:
parallelSentence.add_explicit_empty_tags()
RuleApplicationHypothesis.set_apertium_data_dir(args.apertium_data_dir)
RuleApplicationHypothesis.set_target_language(args.target_language)
print >> sys.stderr, "Loading ATs ..."
ruleList = RuleList()
#load alignment templates
if args.alignment_templates:
if args.alignment_templates.lower().endswith(".gz"):
gfile = gzip.open(args.alignment_templates)
else:
gfile = open(args.alignment_templates)
for line in gfile:
line = line.strip().decode('utf-8')
at = AlignmentTemplate()
at.parse(line)
ruleList.add(at)
gfile.close()
print >> sys.stderr, "... done"
print >> sys.stderr, "Loading sentences ..."
#load sentences
sentences = list()
if args.sentences:
if args.sentences.lower().endswith(".gz"):
gfile = gzip.open(args.sentences)
else:
gfile = open(args.sentences)
for line in gfile:
line = line.strip().decode('utf-8')
parallelSentence = ParallelSentence()
parallelSentence.parse(line,
ruleList=RuleList()
RuleApplicationHypothesis.set_target_language(args.target_language)
RuleApplicationHypothesis.set_apertium_data_dir(args.apertium_data_dir)
RuleApplicationHypothesis.set_minimum_covered_words(args.minimum_covered_words)
#load alignment templates
if args.alignment_templates.endswith(".gz"):
gfile=gzip.open(args.alignment_templates)
else:
gfile=open(args.alignment_templates)
for line in gfile:
line=line.strip().decode('utf-8')
at=AlignmentTemplate()
at.parse(line)
ruleList.add(at)
gfile.close()
ruleLists=[ruleList]
if args.alternative_alignment_templates:
altRuleList=RuleList()
gfile=gzip.open(args.alternative_alignment_templates)
for line in gfile:
line=line.strip().decode('utf-8')
at=AlignmentTemplate()
at.parse(line)
altRuleList.add(at)
gfile.close()
ruleLists=[ruleList,altRuleList]
boxesCoverage=False
RuleApplicationHypothesis.set_target_language(args.target_language)
RuleApplicationHypothesis.set_apertium_data_dir(args.apertium_data_dir)
RuleApplicationHypothesis.set_minimum_covered_words(
args.minimum_covered_words)
#load alignment templates
if args.alignment_templates.endswith(".gz"):
gfile = gzip.open(args.alignment_templates)
else:
gfile = open(args.alignment_templates)
for line in gfile:
line = line.strip().decode('utf-8')
at = AlignmentTemplate()
at.parse(line)
ruleList.add(at)
gfile.close()
ruleLists = [ruleList]
if args.alternative_alignment_templates:
altRuleList = RuleList()
gfile = gzip.open(args.alternative_alignment_templates)
for line in gfile:
line = line.strip().decode('utf-8')
at = AlignmentTemplate()
at.parse(line)
altRuleList.add(at)
gfile.close()
ruleLists = [ruleList, altRuleList]
boxesCoverage = False
