def test_setting_all_short_connectors(self):
po = ParseOptions()
po.all_short_connectors = True
self.assertEqual(po.all_short_connectors, True)
self.assertEqual(clg.parse_options_get_all_short_connectors(po._obj), 1)
po.all_short_connectors = False
self.assertEqual(po.all_short_connectors, False)
self.assertEqual(clg.parse_options_get_all_short_connectors(po._obj), 0)
def test_setting_islands_ok(self):
po = ParseOptions()
po.islands_ok = True
self.assertEqual(po.islands_ok, True)
self.assertEqual(clg.parse_options_get_islands_ok(po._obj), 1)
po.islands_ok = False
self.assertEqual(po.islands_ok, False)
self.assertEqual(clg.parse_options_get_islands_ok(po._obj), 0)
def test_setting_display_morphology(self):
po = ParseOptions()
po.display_morphology = True
self.assertEqual(po.display_morphology, True)
self.assertEqual(clg.parse_options_get_display_morphology(po._obj), 1)
po.display_morphology = False
self.assertEqual(po.display_morphology, False)
self.assertEqual(clg.parse_options_get_display_morphology(po._obj), 0)
def setUpModule():
datadir = os.getenv("LINK_GRAMMAR_DATA", "")
if datadir:
clg.dictionary_set_data_dir(datadir)
clg.test_data_srcdir = os.getenv("srcdir")
if clg.test_data_srcdir:
clg.test_data_srcdir += "/"
else:
clg.test_data_srcdir = ""
def linkage_stat(psent, lang, lkgs, sent_po):
"""
This function mimics the linkage status report style of link-parser
"""
random = ' of {} random linkages'. \
format(clg.sentence_num_linkages_post_processed((psent._obj))) \
if clg.sentence_num_linkages_found(psent._obj) > sent_po.linkage_limit else ''
print ('{}: Found {} linkage{} ({}{} had no P.P. violations)'. \
format(lang, clg.sentence_num_linkages_found(psent._obj),
s(clg.sentence_num_linkages_found(psent._obj)), len(lkgs), random))
def test_setting_verbosity(self):
po = ParseOptions()
po.verbosity = 2
#Ensure that the PO object reports the value correctly
self.assertEqual(po.verbosity, 2)
#Ensure that it's actually setting it.
self.assertEqual(clg.parse_options_get_verbosity(po._obj), 2)
def linkage_testfile(self, lgdict, popt, desc = ''):
"""
Reads sentences and their corresponding
linkage diagrams / constituent printings.
"""
if '' != desc:
desc = desc + '-'
testfile = clg.test_data_srcdir + "parses-" + desc + clg.dictionary_get_lang(lgdict._obj) + ".txt"
parses = open(testfile)
diagram = None
sent = None
lineno = 0
for line in parses:
lineno += 1
# Lines starting with I are the input sentences
if 'I' == line[0]:
sent = line[1:]
diagram = ""
constituents = ""
linkages = Sentence(sent, lgdict, popt).parse()
linkage = linkages.next()
# Generate the next linkage of the last input sentence
if 'N' == line[0]:
diagram = ""
constituents = ""
linkage = next(linkages, None)
if not linkage:
self.assertTrue(linkage, "{}:{}: Sentence has too few linkages".format(testfile, lineno))
# Lines starting with O are the parse diagram
# It ends with an empty line
if 'O' == line[0]:
diagram += line[1:]
if '\n' == line[1] and 1 < len(diagram):
self.assertEqual(linkage.diagram(), diagram)
# Lines starting with C are the constituent output (type 1)
# It ends with an empty line
if 'C' == line[0]:
if '\n' == line[1] and 1 < len(constituents):
self.assertEqual(linkage.constituent_tree(), constituents)
constituents += line[1:]
parses.close()
def test_setting_min_null_count(self):
po = ParseOptions()
po.min_null_count = 3
self.assertEqual(clg.parse_options_get_min_null_count(po._obj), 3)
def test_setting_disjunct_cost(self):
po = ParseOptions()
po.disjunct_cost = 3.0
self.assertEqual(clg.parse_options_get_disjunct_cost(po._obj), 3.0)
def test_setting_linkage_limit(self):
po = ParseOptions()
po.linkage_limit = 3
self.assertEqual(clg.parse_options_get_linkage_limit(po._obj), 3)
def linkage_testfile(self, lgdict, popt, desc = ''):
"""
Reads sentences and their corresponding
linkage diagrams / constituent printings.
"""
self.__class__.longMessage = True
if desc != '':
desc = desc + '-'
testfile = clg.test_data_srcdir + "parses-" + desc + clg.dictionary_get_lang(lgdict._obj) + ".txt"
parses = open(testfile, "rb")
diagram = None
constituents = None
sent = None
lineno = 0
# Function code and file format sanity check
self.opcode_detected = 0
def validate_opcode(ctxt=self, O=False, C=False):
ctxt.opcode_detected += 1
if O:
self.assertFalse(diagram, "at {}:{}: Unfinished diagram entry".format(testfile, lineno))
if C:
self.assertFalse(constituents, "at {}:{}: Unfinished constituents entry".format(testfile, lineno))
for line in parses:
lineno += 1
if sys.version_info > (3, 0):
line = line.decode('utf-8')
# Lines starting with I are the input sentences
if line[0] == 'I':
validate_opcode(O=True, C=True)
sent = line[1:]
diagram = ""
constituents = ""
linkages = Sentence(sent, lgdict, popt).parse()
linkage = next(linkages, None)
self.assertTrue(linkage, "at {}:{}: Sentence has no linkages".format(testfile, lineno))
# Generate the next linkage of the last input sentence
if line[0] == 'N' :
validate_opcode(O=True, C=True)
diagram = ""
constituents = ""
linkage = next(linkages, None)
self.assertTrue(linkage, "at {}:{}: Sentence has too few linkages".format(testfile, lineno))
# Lines starting with O are the parse diagram
# It ends with an empty line
if line[0] == 'O' :
validate_opcode(C=True)
diagram += line[1:]
if line[1] == '\n' and len(diagram) > 1:
self.assertEqual(linkage.diagram(), diagram, "at {}:{}".format(testfile, lineno))
diagram = None
# Lines starting with C are the constituent output (type 1)
# It ends with an empty line
if line[0] == 'C':
validate_opcode(O=True)
if line[1] == '\n' and len(constituents) > 1:
self.assertEqual(linkage.constituent_tree(), constituents, "at {}:{}".format(testfile, lineno))
constituents = None
else:
constituents += line[1:]
parses.close()
validate_opcode(O=True, C=True)
self.assertGreaterEqual(self.opcode_detected, 2, "Nothing has been done for " + testfile)
unittest.TestCase.assertRegexpMatches = unittest.TestCase.assertRegex
import lg_testutils # Found in the same directory of this test script
# Show information on this program run
print('Running by:', sys.executable)
print('Running {} in:'.format(sys.argv[0]), os.getcwd())
for v in 'PYTHONPATH', 'srcdir', 'LINK_GRAMMAR_DATA':
print('{}={}'.format(v, os.environ.get(v)))
#===
from linkgrammar import (Sentence, Linkage, ParseOptions, Link, Dictionary,
LG_Error, LG_DictionaryError, LG_TimerExhausted,
Clinkgrammar as clg)
print(clg.linkgrammar_get_configuration())
# Show the location and version of the bindings modules
for imported_module in 'linkgrammar$', 'clinkgrammar', '_clinkgrammar', 'lg_testutils':
module_found = False
for module in sys.modules:
if re.search(r'^(linkgrammar\.)?' + imported_module, module):
print("Using", sys.modules[module], end='')
if hasattr(sys.modules[module], '__version__'):
print(' version', sys.modules[module].__version__, end='')
print()
module_found = True
if not module_found:
print("Warning: Module", imported_module, "not loaded.")
sys.stdout.flush()
def test_setting_max_null_count(self):
po = ParseOptions()
po.max_null_count = 3
self.assertEqual(clg.parse_options_get_max_null_count(po._obj), 3)
def test_setting_max_parse_time(self):
po = ParseOptions()
po.max_parse_time = 3
self.assertEqual(clg.parse_options_get_max_parse_time(po._obj), 3)
def test_specifying_parse_options(self):
po = ParseOptions(linkage_limit=99)
self.assertEqual(clg.parse_options_get_linkage_limit(po._obj), 99)
def test_setting_linkage_limit(self):
po = ParseOptions()
po.linkage_limit = 3
self.assertEqual(clg.parse_options_get_linkage_limit(po._obj), 3)
def test_setting_short_length(self):
po = ParseOptions()
po.short_length = 3
self.assertEqual(clg.parse_options_get_short_length(po._obj), 3)
def test_setting_short_length(self):
po = ParseOptions()
po.short_length = 3
self.assertEqual(clg.parse_options_get_short_length(po._obj), 3)
#! /usr/bin/env python
# -*- coding: utf8 -*-
"""Link Grammar example usage"""
from __future__ import print_function, division # We require Python 2.6 or later
from linkgrammar import Sentence, ParseOptions, Dictionary, Clinkgrammar as clg
print ("Version:", clg.linkgrammar_get_version())
po = ParseOptions(verbosity=1)
def desc(lkg):
print (lkg.diagram())
print ('Postscript:')
print (lkg.postscript())
print ('---')
def s(q):
return '' if q == 1 else 's'
def linkage_stat(psent, lang, lkgs, sent_po):
"""
This function mimics the linkage status report style of link-parser
"""
random = ' of {} random linkages'. \
format(clg.sentence_num_linkages_post_processed((psent._obj))) \
if clg.sentence_num_linkages_found(psent._obj) > sent_po.linkage_limit else ''
print ('{}: Found {} linkage{} ({}{} had no P.P. violations)'. \
format(lang, clg.sentence_num_linkages_found(psent._obj),
s(clg.sentence_num_linkages_found(psent._obj)), len(lkgs), random))
def test_specifying_parse_options(self):
po = ParseOptions(linkage_limit=99)
self.assertEqual(clg.parse_options_get_linkage_limit(po._obj), 99)
def test_setting_max_parse_time(self):
po = ParseOptions()
po.max_parse_time = 3
self.assertEqual(clg.parse_options_get_max_parse_time(po._obj), 3)
def test_setting_disjunct_cost(self):
po = ParseOptions()
po.disjunct_cost = 3.0
self.assertEqual(clg.parse_options_get_disjunct_cost(po._obj), 3.0)