def build_key2data(keywords, names):
filters = [
featureFilter.Filter(long_time=duration.LONG_TIME, keywords=keywords),
featureFilter.Filter(long_time=digraph.LONG_TIME, keywords=keywords),
featureFilter.Filter(long_time=trigraph.LONG_TIME, keywords=keywords),
featureFilter.Filter(long_time=digraphRatio.LONG_TIME,
keywords=keywords),
]
feature_types = featureGenerator.ALL_TYPES
people = dict()
for name in names:
print("Building " + name)
path = os.path.join(main_directory, name)
events, _ = parseFile.get_events(path,
filters[0],
is_uniform=False,
with_sample=False)
people[name] = dict()
for filter, feature_type in zip(filters, feature_types):
keys2data = featureGenerator.create_data(events, filter,
feature_type)
people[name].update(
keys2data
) #{key:data for key,data in keys2data.items() if key in chosen_keys})
return people
def generate_people(filters, feature_types, is_uniform=False):
default_filter = filters[0] #should all be identical
people = []
for name in os.listdir(main_directory):
print("Building " + name)
path = os.path.join(main_directory, name)
if os.path.isdir(path):
events, _ = parseFile.get_events(path, default_filter, is_uniform, with_sample=False)
people.append(Person(name, events, filters, feature_types, is_uniform))
return people
def build_key2data(keywords, names):
filters = [featureFilter.Filter(long_time=duration.LONG_TIME,keywords=keywords),
featureFilter.Filter(long_time=digraph.LONG_TIME,keywords=keywords),
featureFilter.Filter(long_time=trigraph.LONG_TIME,keywords=keywords),
featureFilter.Filter(long_time=digraphRatio.LONG_TIME,keywords=keywords),
]
feature_types = featureGenerator.ALL_TYPES
people = dict()
for name in names:
print("Building " + name)
path = os.path.join(main_directory, name)
events, _ = parseFile.get_events(path, filters[0], is_uniform=False, with_sample=False)
people[name] = dict()
for filter, feature_type in zip(filters, feature_types):
keys2data = featureGenerator.create_data(events,filter,feature_type)
people[name].update( keys2data )#{key:data for key,data in keys2data.items() if key in chosen_keys})
return people
names = ['Gal Oz-Ari', 'Gil Boazi', 'Nir Yaron', 'Guy Levanon', 'Yonathan Schwammenthal', 'Matan Levine', 'Ohad Ben-Or', 'Dor Aharonson', 'Yuval Itkin', 'Yonatan Caspi', 'Noam Greenberg', 'Adi Asher', 'Yovel Rom']
#filter and get feature arrays for them
keywords = ('java', 'Java', 'Eclipse', 'IntelliJ', 'IDEA')
languages = [featureFilter.ENGLISH]
filters = [featureFilter.Filter(long_time=duration.LONG_TIME,keywords=keywords),
featureFilter.Filter(long_time=digraph.LONG_TIME,keywords=keywords),
featureFilter.Filter(long_time=trigraph.LONG_TIME,keywords=keywords),
featureFilter.Filter(long_time=digraphRatio.LONG_TIME,keywords=keywords),
]
feature_types = featureGenerator.ALL_TYPES
people = dict() #name:[time1, time2, time3...]
for name in names:
print("Building " + name)
path = os.path.join(main_directory, name)
events, _ = parseFile.get_events(path, filters[0], is_uniform=False, with_sample=False)
people[name] = dict()
for filter, feature_type in zip(filters, feature_types):
keys2data = featureGenerator.create_data(events,filter,feature_type)
people[name].update( keys2data )#{key:data for key,data in keys2data.items() if key in chosen_keys})
loc = r"F:\Clouds\Dropbox\SMOP\AnalysisCompressed\eclipse_english_people2key2data.p"
pickle.dump( people, open( loc, "wb+" ) )
#find best features (most common of each type)
def best_keys2(people, feature_type):
#for people from featureGenerator.create_data
keys = dict()
for p in people.values():
for k,li in p.items():
if k[-1] == feature_type:
names = names3
#filter and get feature arrays for them
keywords = tuple() #('java', 'Java', 'Eclipse', 'IntelliJ', 'IDEA')
languages = featureFilter.ALL_LANGUAGES #[featureFilter.ENGLISH]
filters = [featureFilter.Filter(long_time=duration.LONG_TIME,keywords=keywords,languages=languages),
featureFilter.Filter(long_time=digraph.LONG_TIME,keywords=keywords,languages=languages),
featureFilter.Filter(long_time=trigraph.LONG_TIME,keywords=keywords,languages=languages),
featureFilter.Filter(long_time=digraphRatio.LONG_TIME,keywords=keywords,languages=languages),
]
feature_types = featureGenerator.ALL_TYPES
people = dict() #name:[time1, time2, time3...]
for name in names:
print("Building " + name)
path = os.path.join(main_directory, name)
events, _ = parseFile.get_events(path, filters[0], is_uniform=False, with_sample=False)
people[name] = dict()
for filter, feature_type in zip(filters, feature_types):
keys2data = featureGenerator.create_data(events,filter,feature_type)
people[name].update( keys2data )#{key:data for key,data in keys2data.items() if key in chosen_keys})
loc = "F:\Clouds\Dropbox\SMOP\AnalysisCompressed\general_people2key2data.p" #r"F:\Clouds\Dropbox\SMOP\AnalysisCompressed\eclipse_english_people2key2data.p"
pickle.dump( people, open( loc, "wb+" ) )
#find best features (most common of each type)
def best_keys2(people, feature_type):
#for people from featureGenerator.create_data
keys = dict()
for p in people.values():
for k,li in p.items():