def insert_tirp_to_tree(tirp_obj_class_1, tirp_obj_class_0, min_ver_support):
"""
This is a recursive function that gets two branches and combines them
:param tirp_obj_class_1:
:param tirp_obj_class_0:
:param min_ver_support:
:return: combined branch
"""
childs = tirp_obj_class_1.get_childes().copy()
for child in childs:
if type(child) != TIRP:
child_tirp = TIRP()
child_tirp.__dict__.clear()
child_tirp.__dict__.update(child)
else:
child_tirp = child
if not child_tirp.get_exist_in_class_0():
curr_tirp = tirp_obj_class_0
childs_class_0 = tirp_obj_class_0.get_childes()
while curr_tirp.get_tirp_size() + 1 < child_tirp.get_tirp_size(
): #need to go down in the tree until geting to right level to insert
childes_new = list()
for child_class_0 in childs_class_0:
if type(child_class_0) != TIRP:
curr_tirp = TIRP()
curr_tirp.__dict__.clear()
curr_tirp.__dict__.update(child_class_0)
childes_new.append(curr_tirp)
else:
childes_new.append(child_class_0)
childs_class_0 = childes_new
for curr_tirp in childs_class_0:
curr_size = curr_tirp.get_tirp_size()
new_symbols = child_tirp.get_symbols()[:curr_size]
num_of_new_rels = int(curr_size * (curr_size - 1) / 2)
new_rels = child_tirp.get_rels()[:num_of_new_rels]
if curr_tirp.get_symbols() == new_symbols:
if curr_tirp.get_rels() == new_rels:
# if curr_tirp.get_tirp_size()+1 < child_tirp.get_tirp_size() and curr_tirp.get_exist_in_class_0():
childs_class_0 = curr_tirp.get_childes()
break # the right level to insert found
if not curr_tirp.get_exist_in_class_0():
for i in range(0, len(childs_class_0)):
if type(childs_class_0[i]) != TIRP:
tirp = TIRP()
tirp.__dict__.clear()
tirp.__dict__.update(childs_class_0[i])
else:
tirp = childs_class_0[i]
if tirp.get_symbols() == child_tirp.get_symbols(
) and tirp.get_rels() == child_tirp.get_rels():
del childs_class_0[i]
break
child_tirp.set_class_0_properties(min_ver_support)
childs_class_0.append(child_tirp)
tirp_obj_class_0 = insert_tirp_to_tree(child_tirp, tirp_obj_class_0,
min_ver_support)
return tirp_obj_class_0
def marge_trees(data_set_path, dir_path_class_0, dir_path_class_1,
states_by_name, min_ver_support):
"""
This function find tirps that exist only in class 1 and insert them to the right place in the tree
:param data_set_path:
:param states_by_name:
:param min_ver_support:
:return:
"""
root_elements_class_1 = list()
root_elements_class_0 = list()
temp_list = list() # for tirps in class 1 that not have childs
with open(data_set_path + dir_path_class_0 + 'root.txt',
"r") as fr: #load roots from class 0
lines = fr.readlines()
for line in lines:
tirp_obj_class_0 = TIRP()
tirp_obj_class_0.__dict__.clear()
tirp_obj_class_0.__dict__.update(json.loads(line))
root_elements_class_0.append(tirp_obj_class_0)
with open(data_set_path + dir_path_class_1 + 'root.txt',
"r") as fr: #load roots from class 1
lines = fr.readlines()
for line in lines:
tirp_obj_class_1 = TIRP()
tirp_obj_class_1.__dict__.clear()
tirp_obj_class_1.__dict__.update(json.loads(line))
root_elements_class_1.append(tirp_obj_class_1)
for root_element_class_1 in root_elements_class_1:
file_name = states_by_name[root_element_class_1.get_symbols()
[0]] + '.txt'
if not root_element_class_1.get_exist_in_class_0(
): # if the root not exist in class 0
with open(data_set_path + dir_path_class_0 + 'root.txt',
"a") as fr:
r = json.dumps(root_element_class_1,
default=lambda x: x.__dict__)
root_element_class_1.set_class_0_properties(min_ver_support)
fr.write("%s\n" % r)
temp_list.append(
root_element_class_1) # insert it to temp list
if os.path.isfile(data_set_path + dir_path_class_1 +
file_name): # if the root has childs
with open(data_set_path + dir_path_class_1 + file_name,
"r") as fr:
tirp = fr.readlines()
with open(data_set_path + dir_path_class_0 + file_name,
"a") as fs:
fs.writelines(
tirp) # insert all the branch to the tree
else: # if the root exist in class 0
for root_element_class_0 in root_elements_class_0:
if root_element_class_0.get_symbols(
)[0] == root_element_class_1.get_symbols(
)[0]: #find the right root in class 0
root_element_class_0.set_exist_in_class_0(
) # update root elements that were found to true
if len(root_element_class_0.get_childes()) == 0:
if os.path.isfile(data_set_path + dir_path_class_1 +
file_name):
root_element_class_0.set_childes(has_childs=True)
break
if os.path.isfile(data_set_path + dir_path_class_1 +
file_name): # if the root has childs
with open(data_set_path + dir_path_class_1 + file_name,
"r") as fr: #load the branch
tirp_class_1 = fr.readline()
tirp_dict_class_1 = json.loads(tirp_class_1)
tirp_obj_class_1 = TIRP()
tirp_obj_class_1.__dict__.clear()
tirp_obj_class_1.__dict__.update(tirp_dict_class_1)
if os.path.isfile(data_set_path + dir_path_class_0 +
file_name): # if the branch exist in class 0
with open(data_set_path + dir_path_class_0 + file_name,
"r") as fr: #load the branch
tirp_class_0 = fr.readline()
tirp_dict_class_0 = json.loads(tirp_class_0)
tirp_obj_class_0 = TIRP()
tirp_obj_class_0.__dict__.clear()
tirp_obj_class_0.__dict__.update(tirp_dict_class_0)
tirp_obj_class_0.set_exist_in_class_0()
# if os.path.isfile(data_set_path + '/tempChunks1/' + file_name): # if the branch exist in class 1
tirp_obj_class_0 = insert_tirp_to_tree(
tirp_obj_class_1, tirp_obj_class_0,
min_ver_support) #combine the branches
else: # if the branch not exist in class 0 but the root exist
if root_element_class_0.get_symbols(
)[0] == root_element_class_1.get_symbols()[0]:
if len(root_element_class_0.get_childes(
)) > 0 and type(
root_element_class_0.get_childes()[0]) == bool:
root_element_class_0.update_childs(list())
tirp_obj_class_0 = insert_tirp_to_tree(
tirp_obj_class_1, root_element_class_0,
min_ver_support)
else:
if os.path.isfile(data_set_path + dir_path_class_1 +
file_name):
tirp_obj_class_0 = tirp_obj_class_1
else:
tirp_obj_class_0 = root_element_class_1
tirp_obj_class_0_json = json.dumps(
tirp_obj_class_0, default=lambda x: x.__dict__)
with open(data_set_path + dir_path_class_0 + file_name,
"w") as fs:
fs.write(tirp_obj_class_0_json)
os.remove(data_set_path + dir_path_class_0 + 'root.txt')
with open(data_set_path + dir_path_class_0 + 'root.txt', "a") as fr:
for root_element_class_0 in root_elements_class_0:
r = json.dumps(root_element_class_0, default=lambda x: x.__dict__)
fr.write("%s\n" % r)
for root_element_class_1 in temp_list:
r = json.dumps(root_element_class_1, default=lambda x: x.__dict__)
fr.write("%s\n" % r)
def parse_main_index(path, dir_path_class_0, dir_path_class_1, states,
states_by_name, class_name, min_ver_support,
second_class_output_file_name, to_add_entities):
TIRP_per_file.clear()
root_elements = list()
if class_name == 'class_0':
path_to_chuncks = path + '/chunks/'
else:
path_to_chuncks = path + '/chunks1/'
with open(path_to_chuncks + 'main_Index.txt') as fp:
line1 = fp.readline()
while line1:
line2 = fp.readline()
file_name = fp.readline().rstrip()
# ParseOutputFile.parse_states_file()
tirp = ParseOutputFile.parse_TIRP(line1, line2, states, path,
class_name, min_ver_support,
file_name,
second_class_output_file_name,
to_add_entities)
if os.path.isfile(path_to_chuncks + file_name):
tirp.set_childes(has_childs=True)
tirp_with_childs = copy.deepcopy(tirp)
tirp_with_childs = get_sub_tree(tirp_with_childs, states,
states_by_name, path,
class_name, min_ver_support,
file_name, to_add_entities)
tirp_with_childs_json = json.dumps(
tirp_with_childs, default=lambda x: x.__dict__)
if class_name == 'class_0':
path_to_file = path + dir_path_class_0 + file_name
else:
path_to_file = path + dir_path_class_1 + file_name
with open(path_to_file, "w") as fs:
fs.write(tirp_with_childs_json)
else:
if class_name == 'class_0' and second_class_output_file_name != 'File does not exist':
if not to_add_entities: # take the tirp with the entities
dir_path_class_1_new = '/tempChunks1_with_entities/'
else:
dir_path_class_1_new = dir_path_class_1
if os.path.isfile(path + dir_path_class_1_new + file_name):
with open(path + dir_path_class_1_new + file_name,
"r") as fr:
tirp_dict = json.load(fr)
tirp_obj = TIRP()
tirp_obj.__dict__.clear()
tirp_obj.__dict__.update(tirp_dict)
tirp.set_class_1_properties(tirp_obj)
if not to_add_entities:
tirp.set_supporting_instances(list())
tirp.set_supporting_entitie(list())
else:
with open(path + dir_path_class_1_new + 'root.txt',
"r") as fr:
lines = fr.readlines()
for line in lines:
tirp_obj_class_1 = TIRP()
tirp_obj_class_1.__dict__.clear()
tirp_obj_class_1.__dict__.update(
json.loads(line))
if tirp_obj_class_1.get_symbols(
) == tirp.get_symbols():
if tirp_obj_class_1.get_rels == tirp.get_rels(
):
tirp.set_class_1_properties(
tirp_obj_class_1)
if not to_add_entities:
tirp.set_supporting_instances(
list())
tirp.set_supporting_entitie(list())
break
s = json.dumps(tirp, default=lambda x: x.__dict__)
root_elements.append(s)
# root_elements.append(TIRP)
TIRP_name = tirp.get_unique_name()
TIRP_per_file[TIRP_name] = file_name
line1 = fp.readline()
return root_elements
def find_Path_of_tirps(symbols,
rels,
data_set_path,
states,
states_by_name=None,
to_add_entities=None):
try:
relations_dict = {
"<": "before",
"m": "meets",
"o": "overlaps",
"f": "finished by",
"c": "contains",
"=": "equals",
"s": "starts",
"-": 7
}
rels = list(filter(None, rels.split('.')))
tirps_path = []
symbols = list(filter(None, symbols.split('-')))
if not to_add_entities:
file_name = symbols[0] + '.txt'
dir_path = '/tempChunks/'
for r in range(0, len(rels)):
rels[r] = relations_dict[rels[r]]
for i in range(0, len(symbols)):
symbol = states[symbols[i]]
symbols[i] = symbol
else:
file_name = states_by_name[symbols[0]] + '.txt'
dir_path = '/tempChunks_with_entities/'
tirp_size = len(symbols)
if os.path.isfile(data_set_path + dir_path + file_name):
with open(data_set_path + dir_path + file_name, "r") as fr:
tirp_dict = json.load(fr)
tirp_obj = TIRP()
tirp_obj.__dict__.clear()
tirp_obj.__dict__.update(tirp_dict)
tirps_path.append(tirp_obj)
if tirp_size > 1:
childs = tirp_obj.get_childes()
while len(tirps_path) < tirp_size:
for child in childs:
curr_tirp = TIRP()
curr_tirp.__dict__.clear()
curr_tirp.__dict__.update(child)
curr_size = curr_tirp.get_tirp_size()
new_symbols = symbols[:curr_size]
num_of_new_rels = int(curr_size * (curr_size - 1) /
2)
new_rels = rels[:num_of_new_rels]
if curr_tirp.get_symbols() == new_symbols:
if curr_tirp.get_rels() == new_rels:
tirps_path.append(curr_tirp)
childs = curr_tirp.get_childes()
break
return tirps_path
else:
with open(data_set_path + dir_path + 'root.txt', "r") as fr:
roots_from_file = fr.readlines()
for line in roots_from_file:
tirp_dict = json.loads(line)
tirp_obj = TIRP()
tirp_obj.__dict__.clear()
tirp_obj.__dict__.update(tirp_dict)
if tirp_obj.get_symbols() == symbols:
if tirp_obj.get_rels() == rels:
return [tirp_obj]
except Exception as e:
print(e)
def parse_output_file(filename, rel_number, states, path, class_name,
min_ver_support, class_1_tirp_file_name,
to_add_entities):
"""
This function create TIRP list from KarmaLego output file.
Output file structure: [0]TIRP_size [1]symbolNumber-symbolNumber-sym...- [2]rel.rel.rel... [3]mean_duration
[4]mean_offset_from_start [5]mean_offset_from_end [6]vertical_support
[7]mean_horizontal_support [8]entity_id [9][start_time-end_time][10] duration [11]offset_from_start [12]offset_from_end
:param filename:
:param rel_number:
:return: TIRPs list
"""
if not input_validation(filename):
return
if rel_allen_seven is rel_number:
relations_dict = {
"<": "before",
"m": "meets",
"o": "overlaps",
"f": "finished by",
"c": "contains",
"=": "equals",
"s": "starts",
"-": 7
}
else:
print("Wrong number of relations")
return
TIRP_list = []
lines = [line.rstrip('\n') for line in open(filename)]
for i in range(0, len(lines) - 1):
if i % 2 == 1:
continue
line_vector = lines[i].split()
next_line = lines[i + 1]
instances = []
entities = list()
TIRP_size = int(line_vector[0])
symbols = list(filter(None, line_vector[1].split('-')))
if states:
for i in range(0, len(symbols)):
symbol = states[symbols[i]]
symbols[i] = symbol
if TIRP_size > 1:
index = 0
relations = list(filter(None, line_vector[index + 2].split('.')))
for r in range(0, len(relations)):
relations[r] = relations_dict[relations[r]]
else:
relations = list()
index = -1
mean_duration = float(line_vector[index + 3])
mean_offset_from_start = float(line_vector[index + 4])
mean_offset_from_end = float(line_vector[index + 5])
vertical_support = int(line_vector[index + 6])
mean_horizontal_support = float(line_vector[index + 7])
get_supporting_instances(entities,
instances,
line_vector,
symbols,
index=index,
next_line=next_line)
TIRP_obj = TIRP(tirp_size=TIRP_size,
symbols=symbols,
relation=relations,
supporting_instances=instances,
supporting_entities=entities,
vertical_support=vertical_support,
mean_horizontal_support=mean_horizontal_support,
mean_duration=mean_duration,
mean_offset_from_start=mean_offset_from_start,
mean_offset_from_end=mean_offset_from_end,
path=path,
min_vertical_support=min_ver_support)
if class_name == 'class_0':
class_1_tirp = find_tirp_in_class_1(path, TIRP_obj,
class_1_tirp_file_name,
to_add_entities)
TIRP_obj.set_exist_in_class_0()
if class_1_tirp:
if not to_add_entities:
class_1_tirp = find_tirp_in_class_1(
path, TIRP_obj, class_1_tirp_file_name, True)
TIRP_obj.set_class_1_properties(class_1_tirp)
if not to_add_entities:
TIRP_obj.set_supporting_instances(list())
TIRP_obj.set_supporting_entitie(list())
TIRP_list.append(TIRP_obj)
return TIRP_list
def find_tirp_in_class_1(path, class_0_tirp, class_1_tirp_file_name,
to_add_entities):
if to_add_entities:
dir_path = path + '/tempChunks1_with_entities/'
else:
dir_path = path + '/tempChunks1/'
path_to_class_1_tirp = dir_path + class_1_tirp_file_name
if os.path.isfile(path_to_class_1_tirp):
with open(path_to_class_1_tirp, "r") as fr:
tirp_dict = json.load(fr)
class_1_tirp = TIRP()
class_1_tirp.__dict__.clear()
class_1_tirp.__dict__.update(tirp_dict)
class_0_tirp_size = class_0_tirp.get_tirp_size()
class_1_tirp_size = class_1_tirp.get_tirp_size()
found = True
if class_0_tirp_size == 1: #root element
class_1_tirp.set_exist_in_class_0()
with open(path_to_class_1_tirp, "w") as fw:
class_1_tirp_json = json.dumps(
class_1_tirp, default=lambda x: x.__dict__)
fw.write(class_1_tirp_json)
root_elements_class_1 = list()
with open(dir_path + 'root.txt', "r") as fr:
lines = fr.readlines()
for line in lines:
tirp_obj_class_1 = TIRP()
tirp_obj_class_1.__dict__.clear()
tirp_obj_class_1.__dict__.update(json.loads(line))
root_elements_class_1.append(tirp_obj_class_1)
for root_element in root_elements_class_1:
if root_element.get_symbols(
)[0] == class_1_tirp.get_symbols()[0]:
root_element.set_exist_in_class_0()
break
os.remove(dir_path + 'root.txt')
with open(dir_path + 'root.txt', "a") as fr:
for root_element in root_elements_class_1:
r = json.dumps(root_element,
default=lambda x: x.__dict__)
fr.write("%s\n" % r)
return class_1_tirp
else:
father = class_1_tirp
childs = class_1_tirp.get_childes()
while class_1_tirp_size < class_0_tirp_size and len(
childs) > 0 and found:
found = False
for index, child in enumerate(childs):
curr_tirp = TIRP()
curr_tirp.__dict__.clear()
curr_tirp.__dict__.update(child)
curr_size = curr_tirp.get_tirp_size()
new_symbols = class_0_tirp.get_symbols()[:curr_size]
num_of_new_rels = int(curr_size * (curr_size - 1) / 2)
new_rels = class_0_tirp.get_rels()[:num_of_new_rels]
if curr_tirp.get_symbols() == new_symbols:
if curr_tirp.get_rels() == new_rels:
if curr_tirp.get_tirp_size(
) == class_0_tirp.get_tirp_size():
curr_tirp.set_exist_in_class_0()
childs[index] = curr_tirp
father.update_childs(childs)
with open(path_to_class_1_tirp, "w") as fw:
class_1_tirp_json = json.dumps(
class_1_tirp,
default=lambda x: x.__dict__)
fw.write(class_1_tirp_json)
return curr_tirp
else:
father = curr_tirp
childs = curr_tirp.get_childes()
class_1_tirp_size = curr_tirp.get_tirp_size(
)
found = True
break
return None
else:
if class_0_tirp.get_tirp_size() == 1:
root_elements_class_1 = list()
with open(dir_path + 'root.txt', "r") as fr:
lines = fr.readlines()
for line in lines:
tirp_obj_class_1 = TIRP()
tirp_obj_class_1.__dict__.clear()
tirp_obj_class_1.__dict__.update(json.loads(line))
root_elements_class_1.append(tirp_obj_class_1)
for root_element in root_elements_class_1:
if root_element.get_symbols()[0] == class_0_tirp.get_symbols(
)[0]:
root_element.set_exist_in_class_0()
os.remove(dir_path + 'root.txt')
with open(dir_path + 'root.txt', "a") as fr:
for element in root_elements_class_1:
r = json.dumps(element,
default=lambda x: x.__dict__)
fr.write("%s\n" % r)
return root_element
return None
else:
return None
def parse_TIRP(line1, line2, states, path, class_name, min_ver_support,
class_1_tirp_file_name, second_class_output_file_name,
to_add_entities):
relations_dict = {
"<": "before",
"m": "meets",
"o": "overlaps",
"f": "finished by",
"c": "contains",
"=": "equals",
"s": "starts",
"-": 7
}
line_vector = line1.split()
instances = []
entities = list()
TIRP_size = int(line_vector[0])
symbols = list(filter(None, line_vector[1].split('-')))
if states:
for i in range(0, len(symbols)):
symbol = states[symbols[i]]
symbols[i] = symbol
if TIRP_size > 1:
index = 0
relations = list(filter(None, line_vector[index + 2].split('.')))
for r in range(0, len(relations)):
relations[r] = relations_dict[relations[r]]
else:
relations = list()
index = -1
mean_duration = float(line_vector[index + 3])
mean_offset_from_start = float(line_vector[index + 4])
mean_offset_from_end = float(line_vector[index + 5])
vertical_support = int(line_vector[index + 6])
mean_horizontal_support = float(line_vector[index + 7])
get_supporting_instances(entities=entities,
instances=instances,
line_vector=line_vector,
symbols=symbols,
index=index,
next_line=line2)
TIRP_obj = TIRP(tirp_size=TIRP_size,
symbols=symbols,
relation=relations,
supporting_instances=instances,
supporting_entities=entities,
vertical_support=vertical_support,
mean_horizontal_support=mean_horizontal_support,
mean_duration=mean_duration,
mean_offset_from_start=mean_offset_from_start,
mean_offset_from_end=mean_offset_from_end,
path=path,
min_vertical_support=min_ver_support)
if class_name == 'class_0' and second_class_output_file_name != 'File does not exist':
TIRP_obj.set_exist_in_class_0()
class_1_tirp = find_tirp_in_class_1(path, TIRP_obj,
class_1_tirp_file_name,
to_add_entities)
if class_1_tirp:
if not to_add_entities:
class_1_tirp = find_tirp_in_class_1(path, TIRP_obj,
class_1_tirp_file_name,
True)
TIRP_obj.set_class_1_properties(class_1_tirp)
if not to_add_entities:
TIRP_obj.set_supporting_instances(list())
TIRP_obj.set_supporting_entitie(list())
return TIRP_obj