def test_pqa_english_noamb_diled_no_generalization(self):
input_parses = module_path + '/tests/data/POC-English-NoAmb/MST-fixed-manually/'
batch_dir = module_path + '/output/test_grammar_learner_' + str(
UTC())[:10]
prj_dir = batch_dir + '/noamb_pqa_diled_no_generalization/'
if check_dir(prj_dir, create=True, verbose='max'):
outpath = prj_dir
# cp,rp :: (test) corpus_path and reference_path:
cp = module_path + '/data/POC-English-NoAmb/poc_english_noamb_corpus.txt'
rp = input_parses + '/poc-english_noAmb-parses-gold.txt'
kwargs = {
'input_parses': input_parses,
'output_grammar': outpath,
'left_wall': '',
'period': False,
'context': 2,
'word_space': 'discrete',
'dim_reduction': 'none',
'clustering': 'group',
'grammar_rules': 2,
'categories_generalization': 'off',
'rules_generalization': 'off',
'tmpath': module_path + '/tmp/',
'linkage_limit': 1000,
'verbose': 'min'
}
re = learn_grammar(**kwargs)
pa, f1, precision, recall = pqa_meter(re['grammar_file'], outpath, cp,
rp, **kwargs)
self.assertTrue(pa * recall > 0.99,
str(pa) + " * " + str(recall) + " > 0.99")
def test_pqa_turtle_ddrkd_no_generalization(self):
input_parses = module_path + '/tests/data/POC-Turtle/MST-fixed-manually/'
batch_dir = module_path + '/output/test_grammar_learner_' + str(
UTC())[:10]
prj_dir = batch_dir + '/turtle_pqa_ddrkd_no_generalization/'
if check_dir(prj_dir, create=True, verbose='max'):
outpath = prj_dir
# cp,rp :: (test) corpus_path and reference_path:
cp = module_path + '/tests/data/POC-Turtle/poc-turtle-corpus.txt'
rp = input_parses + '/poc-turtle-parses-gold.txt'
kwargs = {
'input_parses': input_parses,
'output_grammar': outpath,
'left_wall': '',
'period': False,
'context': 2,
'word_space': 'vectors',
'dim_reduction': 'svd',
'clustering': ('kmeans', 'kmeans++', 18),
#-'cluster_range' : (2,50,9) ,
'cluster_range': (20, 2, 9),
'grammar_rules': 2,
'categories_generalization': 'off',
'rules_generalization': 'off',
'tmpath': module_path + '/tmp/',
'linkage_limit': 1000,
'verbose': 'min'
}
re = learn_grammar(**kwargs)
pa, f1, precision, recall = pqa_meter(re['grammar_file'], outpath, cp,
rp, **kwargs)
self.assertTrue(pa * recall > 0.99,
str(pa) + " * " + str(recall) + " > 0.99")
def test_pqa_turtle_diled_no_generalization(self):
input_parses = module_path + '/tests/data/POC-Turtle/MST-fixed-manually'
batch_dir = module_path + '/output/test_grammar_learner_' + str(UTC())[:10]
prj_dir = batch_dir + '/turtle_pqa_diled_no_generalization/'
if check_dir(prj_dir, create=True, verbose='max'): outpath = prj_dir
# cp,rp :: (test) corpus_path and reference_path:
cp = module_path + '/tests/data/POC-Turtle/poc-turtle-corpus.txt'
rp = input_parses + '/poc-turtle-parses-gold.txt'
kwargs = {
'input_parses' : input_parses,
'output_grammar': outpath,
'left_wall' : '' ,
'period' : False ,
'context' : 2 ,
'word_space' : 'discrete' ,
'dim_reduction' : 'none' ,
'clustering' : 'group' ,
'grammar_rules' : 2 ,
'categories_generalization' : 'off' ,
'rules_generalization' : 'off' ,
'tmpath' : module_path + '/tmp/',
'linkage_limit' : 1000,
'verbose' : 'min'
}
re = learn_grammar(**kwargs)
# 81019 changes: # FIXME: DEL comments
# a, q, qa = pqa_meter(re['grammar_file'], outpath, cp, rp, **kwargs)
# print('parse-ability, parse-quality:', a, q)
# assert a*q > 0.99
# self.assertTrue(a*q*Decimal("100") > 0.99, str(a) + " * " + str(q) + " * 100 !> 0.99")
pa, f1, precision, recall = pqa_meter(re['grammar_file'], outpath, cp, rp, **kwargs)
# pa, f1, precision, recall: <float> 0.0 - 1.0
self.assertTrue(pa*recall > 0.99, str(pa) + " * " + str(recall) + " > 0.99")
def test_turtle_generalize_rules(self):
base = module_path + '/tests/data/POC-Turtle/' + \
'generalized_rules/dict_6C_2018-10-03_0006.4.0.dict'
input_parses = module_path + '/tests/data/POC-Turtle/MST-fixed-manually/'
batch_dir = module_path + '/output/test_grammar_learner_' + str(UTC())[:10]
prj_dir = batch_dir + '/turtle_lw_&_dot_generalized_rules/'
if check_dir(prj_dir, create=True, verbose='max'):
outpath = prj_dir
kwargs = {
'input_parses' : input_parses,
'output_grammar': outpath,
'left_wall' : 'LEFT-WALL' ,
'period' : True ,
'context' : 2 ,
'word_space' : 'discrete' ,
'dim_reduction' : 'none' ,
'clustering' : 'group' ,
'grammar_rules' : 2 ,
'categories_generalization' : 'off' ,
'rules_generalization' : 'jaccard',
'tmpath' : module_path + '/tmp/',
'verbose' : 'none'
}
response = learn_grammar(**kwargs)
with open(response['grammar_file'], 'r') as f:
rules = f.read().splitlines()
rule_list = [line for line in rules if line[0:1] in ['"', '(']]
with open(base, 'r') as f: lst = f.read().splitlines()
base_list = [line for line in lst if line[0:1] in ['"', '(']]
if len(rule_list) == len(base_list):
assert rule_list == base_list
else: assert len(rule_list) == len(base_list)
def test_pqa_english_noamb_ddrkd_no_generalization(self):
input_parses = module_path + '/tests/data/POC-English-NoAmb/MST-fixed-manually/'
batch_dir = module_path + '/output/test_grammar_learner_' + str(
UTC())[:10]
prj_dir = batch_dir + '/noamb_pqa_ddrkd_no_generalization/'
if check_dir(prj_dir, create=True, verbose='max'):
outpath = prj_dir
# cp,rp :: (test) corpus_path and reference_path:
cp = module_path + '/data/POC-English-NoAmb/poc_english_noamb_corpus.txt'
rp = input_parses + '/poc-english_noAmb-parses-gold.txt'
kwargs = {
'input_parses': input_parses,
'output_grammar': outpath,
'left_wall': '',
'period': False,
'context': 2,
'word_space': 'vectors',
'dim_reduction': 'svd',
'clustering': ('kmeans', 'kmeans++', 18),
'cluster_range': (12, 12, 5),
'grammar_rules': 2,
'categories_generalization': 'off',
'rules_generalization': 'off',
'tmpath': module_path + '/tmp/',
'linkage_limit': 1000,
'verbose': 'min'
}
# Sometimes pqa_meter(with test_grammar updated 2018-10-19) returns pa,recall = 0,0
# FIXME: check with further test_grammar updates and delete.
x = 0.
n = 0
while x < 0.1:
re = learn_grammar(**kwargs)
pa, f1, precision, recall = pqa_meter(re['grammar_file'], outpath,
cp, rp, **kwargs)
print(
f'\nnoAmb dDRKd: pa {round(pa,3)}, f1 {round(f1,3)}, precision {round(precision,3)}, recall {round(recall,3)} \n'
)
x = pa * recall
n += 1
if n > 24: break
self.assertTrue(pa * recall > 0.99,
str(pa) + " * " + str(recall) + " > 0.99")
1.0, # level = 0, 1, 0.-0.99..: 0 - max number of clusters
'categories_generalization':
'off', # 'off' / 'cosine' - cosine similarity, 'jaccard'
'categories_merge':
0.8, # merge categories with similarity > this 'merge' criteria
'categories_aggregation':
0.2, # aggregate categories with similarity > this criteria
'grammar_rules':
2, # 1: 'connectors' / 2 - 'disjuncts' / 0 - 'words' (TODO?)
'rules_generalization':
'off', # 'off' / 'cosine' - cosine similarity, 'jaccard'
'rules_merge':
0.8, # merge rules with similarity > this 'merge' criteria
'rules_aggregation':
0.2, # aggregate rules similarity > this criteria
'tmpath': module_path + '/tmp/', # legacy, default if not temp_dir
'verbose':
'max', # display intermediate results: 'none', 'min', 'mid', 'max'
'linkage_limit': 10000, # Link Grammar parameter for tests'
'add_disjunct_costs':
True, # add disjunct costs when saving grammar rules to LG dictionary file
'disjunct_cost_function': 'reverse_count' # 1/disjunct_count
}
#print(kwargs)
response = learn_grammar(**kwargs)
with open(response['grammar_file'], 'r') as f:
rules = f.read().splitlines()
rule_list = [line for line in rules if line[0:1] in ['"', '(']]
except:
print("--ERR--", v, "has no links")
