def __generate_single_bug(self, bug_id, drop_rate):
"""Generate an incomplete bug report text.
Args:
bug_id: int, original bug id.
drop_rate: float, 0.0 for not drop, 1.0 for totally drop.
Returns:
IRReport
"""
from ir_text import IRText
from ir_term_count import IRTermCount
from ir_report import IRReport
# get description and summary
summary, description = IRText.get_summary_and_description_of_bug(bug_id)
create_ts, product = IRText.get_basic_info_of_bug(bug_id)
if drop_rate > 0.001:
summary, description = \
IRTermCount.create_incomplete_report(summary, description, drop_rate)
print description
new_report = IRReport(summary, description)
new_report.set_stacktrace(IRText.get_stacktrace_of_bug(bug_id))
new_report.set_dummy_bug_id(bug_id)
new_report.set_basic_info(create_ts, product)
return new_report
def test_compare_stackinfo(self):
from ir_log import IRLog
from ir_config import IRConfig
from ir_mongodb_helper import IRMongodbHelper
from ir_gnome_st_tools import IRSTTools
from ir_text import IRText
from random import randint
import pymongo
IRLog.get_instance().start_log()
IRConfig.get_instance().load('../data/test/bug_test.cfg')
IRText.parse_info_level1('../data/test/stacktrace_test')
con = IRMongodbHelper.get_instance().get_connection()
db = con[IRConfig.get_instance().get('bug_db_name')]
assert None != db
col = db[IRConfig.get_instance().get('bug_text_collection_name')]
assert None != col
bugs = col.find()
total = col.count()
st1 = bugs[0]["stacktrace"]
for i in range(total):
st2 = bugs[i]["stacktrace"]
result_weight = IRSTTools.compare_stackinfo(st1, st2, 'weight')
result_max = IRSTTools.compare_stackinfo(st1, st2, 'max')
IRLog.get_instance().println('Weight: %f, Max: %f' \
% (result_weight, result_max))
IRLog.get_instance().stop_log()
def test_compare_stackinfo(self):
from ir_log import IRLog
from ir_config import IRConfig
from ir_mongodb_helper import IRMongodbHelper
from ir_gnome_st_tools import IRSTTools
from ir_text import IRText
from random import randint
import pymongo
IRLog.get_instance().start_log()
IRConfig.get_instance().load('../data/test/bug_test.cfg')
IRText.parse_info_level1('../data/test/stacktrace_test')
con = IRMongodbHelper.get_instance().get_connection()
db = con[IRConfig.get_instance().get('bug_db_name')]
assert None != db
col = db[IRConfig.get_instance().get('bug_text_collection_name')]
assert None != col
bugs = col.find()
total = col.count()
st1 = bugs[0]["stacktrace"]
for i in range(total):
st2 = bugs[i]["stacktrace"]
result_weight = IRSTTools.compare_stackinfo(st1, st2, 'weight')
result_max = IRSTTools.compare_stackinfo(st1, st2, 'max')
IRLog.get_instance().println('Weight: %f, Max: %f' \
% (result_weight, result_max))
IRLog.get_instance().stop_log()
def test_filter(self):
from ir_log import IRLog
from ir_config import IRConfig
from ir_mongodb_helper import IRMongodbHelper
from ir_gnome_st_tools import IRSTTools
from ir_text import IRText
import pymongo
IRLog.get_instance().start_log()
IRConfig.get_instance().load('../data/test/bug_test.cfg')
IRText.parse_info_level1('../data/test/info_level1_test')
con = IRMongodbHelper.get_instance().get_connection()
db = con[IRConfig.get_instance().get('bug_db_name')]
assert None != db
col = db[IRConfig.get_instance().get('bug_text_collection_name')]
assert None != col
# Maybe a bug here:
# The test of filter (originally) depends on parse_info_level1
# But parse_info_level1 seems to invoke filter...
for bug in col.find():
# TODO: it's not correct. no stacktrace in desc
desc, stack = IRSTTools.filter(bug["desc"])
IRLog.get_instance().stop_log()
def test_filter(self):
from ir_log import IRLog
from ir_config import IRConfig
from ir_mongodb_helper import IRMongodbHelper
from ir_gnome_st_tools import IRSTTools
from ir_text import IRText
import pymongo
IRLog.get_instance().start_log()
IRConfig.get_instance().load('../data/test/bug_test.cfg')
IRText.parse_info_level1('../data/test/info_level1_test')
con = IRMongodbHelper.get_instance().get_connection()
db = con[IRConfig.get_instance().get('bug_db_name')]
assert None != db
col = db[IRConfig.get_instance().get('bug_text_collection_name')]
assert None != col
# Maybe a bug here:
# The test of filter (originally) depends on parse_info_level1
# But parse_info_level1 seems to invoke filter...
for bug in col.find():
# TODO: it's not correct. no stacktrace in desc
desc, stack = IRSTTools.filter(bug["desc"])
IRLog.get_instance().stop_log()
def __generate_single_bug(self, bug_id, drop_rate):
"""Generate an incomplete bug report text.
Args:
bug_id: int, original bug id.
drop_rate: float, 0.0 for not drop, 1.0 for totally drop.
Returns:
IRReport
"""
from ir_text import IRText
from ir_term_count import IRTermCount
from ir_report import IRReport
# get description and summary
summary, description = IRText.get_summary_and_description_of_bug(
bug_id)
create_ts, product = IRText.get_basic_info_of_bug(bug_id)
if drop_rate > 0.001:
summary, description = \
IRTermCount.create_incomplete_report(summary, description, drop_rate)
print description
new_report = IRReport(summary, description)
new_report.set_stacktrace(IRText.get_stacktrace_of_bug(bug_id))
new_report.set_dummy_bug_id(bug_id)
new_report.set_basic_info(create_ts, product)
return new_report
def test_parse_info_level1(self):
#import sys
#sys.path.append('../bin/')
from ir_log import IRLog
from ir_text import IRText
from ir_config import IRConfig
from ir_mongodb_helper import IRMongodbHelper
IRLog.get_instance().start_log()
IRConfig.get_instance().load('../data/test/bug_test.cfg')
assert None != IRConfig.get_instance()
IRText.parse_info_level1('../data/test/info_level1_test')
IRLog.get_instance().stop_log()
con = IRMongodbHelper.get_instance().get_connection()
db = con[IRConfig.get_instance().get('bug_db_name')]
assert None != db
col = db[IRConfig.get_instance().get('bug_text_collection_name')]
assert None != col
# in the test data, we have 1000 in total.
# within, 40 have no resolution, 154 are incomplete
assert 833 == col.count()
assert 'gnome is full of bugs ! (100000 currently)' == \
col.find({'bug_id':100000})[0]["summ"]
res = col.find(
{"summ": {
'$regex': '(>)|(<)|(")|(&apo)s|(&)'
}})
assert res.count() == 0
def test_parse_info_level1(self):
#import sys
#sys.path.append('../bin/')
from ir_log import IRLog
from ir_text import IRText
from ir_config import IRConfig
from ir_mongodb_helper import IRMongodbHelper
IRLog.get_instance().start_log()
IRConfig.get_instance().load('../data/test/bug_test.cfg')
assert None != IRConfig.get_instance()
IRText.parse_info_level1('../data/test/info_level1_test')
IRLog.get_instance().stop_log()
con = IRMongodbHelper.get_instance().get_connection()
db = con[IRConfig.get_instance().get('bug_db_name')]
assert None != db
col = db[IRConfig.get_instance().get('bug_text_collection_name')]
assert None != col
# in the test data, we have 1000 in total.
# within, 40 have no resolution, 154 are incomplete
assert 833 == col.count()
assert 'gnome is full of bugs ! (100000 currently)' == \
col.find({'bug_id':100000})[0]["summ"]
res = col.find({"summ":{'$regex':'(>)|(<)|(")|(&apo)s|(&)'}})
assert res.count() == 0
def server_cache(msg, res):
from ir_log import IRLog
from ir_text import IRText
from ir_tfidf import IRTFIDF
from ir_document_count import IRDocumentCount
IRLog.get_instance().println('Server is caching data')
IRText.cache_all_data()
IRTFIDF.cache_all_data()
IRDocumentCount.cache_all_data()
IRLog.get_instance().println('Server cached data')
return SIGNAL_CONTINUE
def server_cache(msg, res):
from ir_log import IRLog
from ir_text import IRText
from ir_tfidf import IRTFIDF
from ir_document_count import IRDocumentCount
IRLog.get_instance().println('Server is caching data')
IRText.cache_all_data()
IRTFIDF.cache_all_data()
IRDocumentCount.cache_all_data()
IRLog.get_instance().println('Server cached data')
return SIGNAL_CONTINUE
def do_remove_bad_reports(cls, config_file):
from ir_log import IRLog
from ir_log import IRProgressBar
from ir_config import IRConfig
import ir_mongodb_helper
from ir_text import IRText
config = IRConfig.get_instance()
config.load(config_file)
bug_id_name = config.get('bug_id_name')
bug_description_name = config.get('bug_description_name')
text_cursor = IRText.get_iterator(None)
remove_ids = []
def iter_text(item):
if IRText.is_drop_report(item[bug_description_name]):
remove_ids.append(item[bug_id_name])
IRLog.get_instance().println('Remove report#=%d' % item[bug_id_name], 3)
IRProgressBar.execute_iteration_for_cursor(text_cursor, iter_text)
# remove from all database
def remove_from_collection(collection_cfg_name):
collection =ir_mongodb_helper.IRCollection( \
'bug_db_name', collection_cfg_name, 'a')
collection.remove({'bug_id':{'$in':remove_ids}})
collection.close()
remove_from_collection('bug_text_collection_name')
remove_from_collection('bug_tfidf_collection_name')
remove_from_collection('bug_duplicate_collection_name')
def test_get_stacktrace_text_of_bug(self):
from ir_log import IRLog
from ir_config import IRConfig
from ir_text import IRText
IRConfig.get_instance().load('../data/test/bug_test.cfg')
stacktrace_text = IRText.get_stacktrace_text_of_bug(104400)
IRLog.get_instance().println('stacktrace_text: %s' % (stacktrace_text))
def similarity_over_all(self):
"""Calculate similarity between bug (summary, description) over
all.
Returns:
dict, {bug_id -> [score, summary_score, description_score, stacktrace_score]}
"""
from ir_log import IRLog
from ir_config import IRConfig
from ir_mongodb_helper import IRCollection
from ir_text import IRText
from ir_tfidf import IRTFIDF
logger = IRLog.get_instance()
search_time_span = 2 * 3600 * 24 * 365
bug_id_name = IRConfig.get_instance().get('bug_id_name')
create_ts_name = IRConfig.get_instance().get('bug_create_ts_name')
product_name = IRConfig.get_instance().get('bug_product_name')
basic_collection = IRCollection('bug_db_name',
'bug_basic_collection_name', 'r')
reports2scan = basic_collection.find({
product_name: self.get_product(),
create_ts_name: {
'$gt': self.get_create_ts() - search_time_span
},
bug_id_name: {
'$nin': self.__exclude_report_ids
}
})
result = {}
IRLog.get_instance().println('Comparing with %d reports.' \
% (reports2scan.count()) )
print self.__summary_text
print self.__description_text
for report in reports2scan:
bug_id = report[bug_id_name]
if bug_id == self.get_dummy_bug_id():
continue
# because we don't want to load stacktrace in case of self.__stacktrace
# being none, we create and fill the info of report manually
other_report = IRReport("", "")
other_report.__summary_tfidf, other_report.__description_tfidf = \
IRTFIDF.get_tfidf_of_bug(bug_id)
# if self.__stacktrace is empty, we don't need to do this
if self.get_stacktrace() is not None and \
self.get_stacktrace().__len__() > 0:
other_report.__stacktrace = IRText.get_stacktrace_of_bug(
bug_id)
if other_report.__stacktrace is None:
other_report.__stacktrace = []
result[bug_id] = self.similarity_with(other_report)
return result
def batch_generate_term_count(cls):
"""Generate term count for text in mongodb database,
and store to database.
"""
from ir_log import IRProgressBar
from ir_text import IRText
from ir_config import IRConfig
from ir_mongodb_helper import IRCollection
# config
bug_id_name = IRConfig.get_instance().get('bug_id_name', 'bug_id')
summary_name = IRConfig.get_instance().get('bug_summary_name', 'summ')
description_name = IRConfig.get_instance().\
get('bug_description_name', 'desc')
termcount_collection = IRCollection(
'bug_db_name', 'bug_termcount_collection_name', 'w')
def iter_text(bug):
summary_bow, description_bow = cls.calculate_term_count(
bug[summary_name], bug[description_name])
termcount_collection.insert({
bug_id_name : bug[bug_id_name],
summary_name : summary_bow,
description_name : description_bow })
IRProgressBar.execute_iteration_for_cursor(IRText.get_iterator({}),
iter_text, "From Text to Term Count")
termcount_collection.create_index([(bug_id_name, IRCollection.ASCENDING)])
termcount_collection.close()
def test_get_stacktrace_text_of_bug(self):
from ir_log import IRLog
from ir_config import IRConfig
from ir_text import IRText
IRConfig.get_instance().load('../data/test/bug_test.cfg')
stacktrace_text = IRText.get_stacktrace_text_of_bug(104400)
IRLog.get_instance().println('stacktrace_text: %s' % (stacktrace_text))
def test_get_summary_and_description_of_bug(self):
from ir_log import IRLog
from ir_config import IRConfig
from ir_text import IRText
IRConfig.get_instance().load('../data/test/bug_test.cfg')
summary, description = IRText.get_summary_and_description_of_bug(100000)
IRLog.get_instance().println('summary: %s' % (summary))
IRLog.get_instance().println('description: %s' % (description))
def test_parse_dump_file(self):
from ir_log import IRLog
from ir_text import IRText
from ir_config import IRConfig
from ir_mongodb_helper import IRMongodbHelper
IRLog.get_instance().start_log()
IRConfig.get_instance().load('../data/test/bug_test.cfg')
assert None != IRConfig.get_instance()
IRText.parse_dump_file('../data/test/dump_file_test')
IRLog.get_instance().stop_log()
con = IRMongodbHelper.get_instance().get_connection()
db = con[IRConfig.get_instance().get('bug_db_name')]
assert None != db
col = db[IRConfig.get_instance().get('bug_text_collection_name')]
assert None != col
assert 1000 == col.count()
def test_parse_dump_file(self):
from ir_log import IRLog
from ir_text import IRText
from ir_config import IRConfig
from ir_mongodb_helper import IRMongodbHelper
IRLog.get_instance().start_log()
IRConfig.get_instance().load('../data/test/bug_test.cfg')
assert None != IRConfig.get_instance()
IRText.parse_dump_file('../data/test/dump_file_test')
IRLog.get_instance().stop_log()
con = IRMongodbHelper.get_instance().get_connection()
db = con[IRConfig.get_instance().get('bug_db_name')]
assert None != db
col = db[IRConfig.get_instance().get('bug_text_collection_name')]
assert None != col
assert 1000 == col.count()
def test_get_summary_and_description_of_bug(self):
from ir_log import IRLog
from ir_config import IRConfig
from ir_text import IRText
IRConfig.get_instance().load('../data/test/bug_test.cfg')
summary, description = IRText.get_summary_and_description_of_bug(
100000)
IRLog.get_instance().println('summary: %s' % (summary))
IRLog.get_instance().println('description: %s' % (description))
def similarity_over_all(self):
"""Calculate similarity between bug (summary, description) over
all.
Returns:
dict, {bug_id -> [score, summary_score, description_score, stacktrace_score]}
"""
from ir_log import IRLog
from ir_config import IRConfig
from ir_mongodb_helper import IRCollection
from ir_text import IRText
from ir_tfidf import IRTFIDF
logger = IRLog.get_instance()
search_time_span = 2 * 3600 * 24 * 365
bug_id_name = IRConfig.get_instance().get('bug_id_name')
create_ts_name = IRConfig.get_instance().get('bug_create_ts_name')
product_name = IRConfig.get_instance().get('bug_product_name')
basic_collection = IRCollection(
'bug_db_name', 'bug_basic_collection_name', 'r')
reports2scan = basic_collection.find({
product_name : self.get_product(),
create_ts_name : {'$gt' : self.get_create_ts() - search_time_span},
bug_id_name : {'$nin' : self.__exclude_report_ids} })
result = {}
IRLog.get_instance().println('Comparing with %d reports.' \
% (reports2scan.count()) )
print self.__summary_text
print self.__description_text
for report in reports2scan:
bug_id = report[bug_id_name]
if bug_id == self.get_dummy_bug_id():
continue
# because we don't want to load stacktrace in case of self.__stacktrace
# being none, we create and fill the info of report manually
other_report = IRReport("", "")
other_report.__summary_tfidf, other_report.__description_tfidf = \
IRTFIDF.get_tfidf_of_bug(bug_id)
# if self.__stacktrace is empty, we don't need to do this
if self.get_stacktrace() is not None and \
self.get_stacktrace().__len__() > 0:
other_report.__stacktrace = IRText.get_stacktrace_of_bug(bug_id)
if other_report.__stacktrace is None:
other_report.__stacktrace = []
result[bug_id] = self.similarity_with(other_report)
return result
def get_stacktrace(self):
if self.__bug_id is None:
return self.__stacktrace
else:
if self.__allow_cache and self.__stacktrace is not None:
return self.__stacktrace
from ir_text import IRText
stack = IRText.get_stacktrace_of_bug(self.__bug_id)
if self.__allow_cache:
self.__stacktrace = stack
return stack
def get_stacktrace(self):
if self.__bug_id is None:
return self.__stacktrace
else:
if self.__allow_cache and self.__stacktrace is not None:
return self.__stacktrace
from ir_text import IRText
stack = IRText.get_stacktrace_of_bug(self.__bug_id)
if self.__allow_cache:
self.__stacktrace = stack
return stack
def get_basic_info(self):
if self.__bug_id is None:
return self.__create_ts, self.__product
else:
if self.__allow_cache and \
self.__create_ts is not None and \
self.__product is not None:
return self.__create_ts, self.__product
from ir_text import IRText
create_ts, product = IRText.get_basic_info_of_bug(self.__bug_id)
if self.__allow_cache:
self.__create_ts, self.__product = (create_ts, product)
return create_ts, product
def get_basic_info(self):
if self.__bug_id is None:
return self.__create_ts, self.__product
else:
if self.__allow_cache and \
self.__create_ts is not None and \
self.__product is not None:
return self.__create_ts, self.__product
from ir_text import IRText
create_ts, product = IRText.get_basic_info_of_bug(self.__bug_id)
if self.__allow_cache:
self.__create_ts, self.__product = (create_ts, product)
return create_ts, product
def get_summary_and_description_text(self):
if self.__bug_id is None:
return [self.__summary_text, self.__description_text]
else:
if self.__allow_cache and \
self.__summary_text is not None and \
self.__description_text is not None:
return [self.__summary_text, self.__description_text]
from ir_text import IRText
summary, description = \
IRText.get_summary_and_description_of_bug(self.__bug_id)
if self.__allow_cache:
self.__summary_text, self.__description_text = \
summary, description
return summary, description
def get_summary_and_description_text(self):
if self.__bug_id is None:
return [self.__summary_text, self.__description_text]
else:
if self.__allow_cache and \
self.__summary_text is not None and \
self.__description_text is not None:
return [self.__summary_text, self.__description_text]
from ir_text import IRText
summary, description = \
IRText.get_summary_and_description_of_bug(self.__bug_id)
if self.__allow_cache:
self.__summary_text, self.__description_text = \
summary, description
return summary, description
def do_test_over_file(self, filename):
"""Do test over the file.
Args:
filename: str, the input file which generated by
generate_incomplete_test_file.
"""
from ir_log import IRLog
from ir_config import IRConfig
from ir_duplicate_group import IRDuplicateGroup
from ir_text import IRText
from ir_term_count import IRTermCount
from ir_tfidf import IRTFIDF
from ir_report import IRReport
from ir_document_count import IRDocumentCount
IRText.cache_all_data()
IRTFIDF.cache_all_data()
IRDocumentCount.cache_all_data()
remove_self_bug_id = IRConfig.get_instance().get_bool('remove_self_bug_id', True)
sim_tot_precision = 0.0
sim_tot_recall = 0.0
sim_bi_tot_recall = 0.0
sim_tot_size = 0
dup_tot_precision = 0.0
dup_tot_recall = 0.0
dup_bi_toto_recall = 0.0
dup_num = 0
test_num = 0
infile = open(filename, 'r')
for line in infile:
IRLog.get_instance().println('----test----')
test_num += 1
line.strip()
new_report = IRReport.from_string(line)
ori_report = IRReport(new_report.get_dummy_bug_id())
#IRLog.get_instance().println('Summary')
#IRTermCount.show_dict_compare(ori_report.get_summary_termcount(),
# new_report.get_summary_termcount())
#IRLog.get_instance().println('Description')
#IRTermCount.show_dict_compare(ori_report.get_description_termcount(),
# new_report.get_description_termcount())
# do test for single
similarities, duplicates = new_report.similarities_and_duplicates()
sim_ids = [sim[0] for sim in similarities]
dup_ids = [dup[0] for dup in duplicates]
IRLog.get_instance().println('Sim ids: %s' % str(sim_ids))
IRLog.get_instance().println('Dup ids: %s' % str(dup_ids))
# evaluate sim
sim_hit, sim_nothit, real_duplicates = \
IRDuplicateGroup.is_in_same_duplicate_group(
new_report.get_dummy_bug_id(), sim_ids, remove_self_bug_id)
# some group contain only one
if real_duplicates.__len__() == 0:
test_num -= 1
continue
precision, recall = self.__report_result(
new_report.get_dummy_bug_id(), sim_hit, sim_nothit, real_duplicates)
sim_tot_precision += precision
sim_tot_recall += recall
sim_tot_size += sim_ids.__len__()
sim_bi_tot_recall += 1 if recall > 0.0 else 0
if dup_ids.__len__() > 0:
dup_num += 1
dup_hit, dup_nothit, real_duplicates = \
IRDuplicateGroup.is_in_same_duplicate_group(
new_report.get_dummy_bug_id(), dup_ids, remove_self_bug_id)
precision, recall = self.__report_result(
new_report.get_dummy_bug_id(), dup_hit, dup_nothit, real_duplicates)
dup_tot_precision += precision
dup_tot_recall += recall
dup_bi_toto_recall += 1 if recall > 0.0 else 0
# general conclusion
if dup_num == 0:
dup_num = 1.0
IRLog.get_instance().println(','.join(['#cases', 'sim pre', 'sim rec', 'sim birec', 'sim size',\
'#dup', 'dup pre', 'dup rec', 'dup birec']))
IRLog.get_instance().println(','.join([str(test_num), \
str(sim_tot_precision/test_num), str(sim_tot_recall/test_num), str(sim_bi_tot_recall/test_num), str(float(sim_tot_size)/test_num), \
str(dup_num), \
str(dup_tot_precision/dup_num), str(dup_tot_recall/dup_num), str(dup_bi_toto_recall/dup_num)]))
infile.close()
def test_cache_all_data(self):
from ir_log import IRLog
from ir_config import IRConfig
from ir_text import IRText
IRConfig.get_instance().load('../data/test/bug_test.cfg')
IRText.cache_all_data()
def do_test_over_file(self, filename):
"""Do test over the file.
Args:
filename: str, the input file which generated by
generate_incomplete_test_file.
"""
from ir_log import IRLog
from ir_config import IRConfig
from ir_duplicate_group import IRDuplicateGroup
from ir_text import IRText
from ir_term_count import IRTermCount
from ir_tfidf import IRTFIDF
from ir_report import IRReport
from ir_document_count import IRDocumentCount
IRText.cache_all_data()
IRTFIDF.cache_all_data()
IRDocumentCount.cache_all_data()
remove_self_bug_id = IRConfig.get_instance().get_bool(
'remove_self_bug_id', True)
sim_tot_precision = 0.0
sim_tot_recall = 0.0
sim_bi_tot_recall = 0.0
sim_tot_size = 0
dup_tot_precision = 0.0
dup_tot_recall = 0.0
dup_bi_toto_recall = 0.0
dup_num = 0
test_num = 0
infile = open(filename, 'r')
for line in infile:
IRLog.get_instance().println('----test----')
test_num += 1
line.strip()
new_report = IRReport.from_string(line)
ori_report = IRReport(new_report.get_dummy_bug_id())
#IRLog.get_instance().println('Summary')
#IRTermCount.show_dict_compare(ori_report.get_summary_termcount(),
# new_report.get_summary_termcount())
#IRLog.get_instance().println('Description')
#IRTermCount.show_dict_compare(ori_report.get_description_termcount(),
# new_report.get_description_termcount())
# do test for single
similarities, duplicates = new_report.similarities_and_duplicates()
sim_ids = [sim[0] for sim in similarities]
dup_ids = [dup[0] for dup in duplicates]
IRLog.get_instance().println('Sim ids: %s' % str(sim_ids))
IRLog.get_instance().println('Dup ids: %s' % str(dup_ids))
# evaluate sim
sim_hit, sim_nothit, real_duplicates = \
IRDuplicateGroup.is_in_same_duplicate_group(
new_report.get_dummy_bug_id(), sim_ids, remove_self_bug_id)
# some group contain only one
if real_duplicates.__len__() == 0:
test_num -= 1
continue
precision, recall = self.__report_result(
new_report.get_dummy_bug_id(), sim_hit, sim_nothit,
real_duplicates)
sim_tot_precision += precision
sim_tot_recall += recall
sim_tot_size += sim_ids.__len__()
sim_bi_tot_recall += 1 if recall > 0.0 else 0
if dup_ids.__len__() > 0:
dup_num += 1
dup_hit, dup_nothit, real_duplicates = \
IRDuplicateGroup.is_in_same_duplicate_group(
new_report.get_dummy_bug_id(), dup_ids, remove_self_bug_id)
precision, recall = self.__report_result(
new_report.get_dummy_bug_id(), dup_hit, dup_nothit,
real_duplicates)
dup_tot_precision += precision
dup_tot_recall += recall
dup_bi_toto_recall += 1 if recall > 0.0 else 0
# general conclusion
if dup_num == 0:
dup_num = 1.0
IRLog.get_instance().println(','.join(['#cases', 'sim pre', 'sim rec', 'sim birec', 'sim size',\
'#dup', 'dup pre', 'dup rec', 'dup birec']))
IRLog.get_instance().println(','.join([str(test_num), \
str(sim_tot_precision/test_num), str(sim_tot_recall/test_num), str(sim_bi_tot_recall/test_num), str(float(sim_tot_size)/test_num), \
str(dup_num), \
str(dup_tot_precision/dup_num), str(dup_tot_recall/dup_num), str(dup_bi_toto_recall/dup_num)]))
infile.close()
def test_cache_all_data(self):
from ir_log import IRLog
from ir_config import IRConfig
from ir_text import IRText
IRConfig.get_instance().load('../data/test/bug_test.cfg')
IRText.cache_all_data()
def iter_text(item):
if IRText.is_drop_report(item[bug_description_name]):
remove_ids.append(item[bug_id_name])
IRLog.get_instance().println('Remove report#=%d' % item[bug_id_name], 3)