AutoRestTest-TrackA / tools /morest /model /operation_dependency_graph.py
minhhungg's picture
feat: initial upload for AutoRestTest Track A datasets
6c50d1f
Raw
History Blame Contribute Delete
10.7 kB
import json
import re
import numpy as np
from graphviz import Digraph
from .sequence import Sequence
from .sequence import SequenceOrigin
class Path:
def __init__(self):
self.nodes = []
def duplicate(self):
path = Path()
path.nodes = list(self.nodes)
return path
def append(self, elem):
self.nodes.append(elem)
def __iter__(self):
return self.nodes.__iter__()
def __getitem__(self, item):
return self.nodes[item]
def __str__(self):
if len(self.nodes):
return ''
return ' -> '.join(self.nodes)
def __hash__(self):
return hash(self.__str__())
def remove(self, elem):
self.nodes.remove(elem)
def __len__(self):
return len(self.nodes)
def __eq__(self, other):
if len(other.nodes) != len(self.nodes) or not isinstance(other, Path):
return False
return str(other) == str(self)
def pop(self, ind):
self.nodes.pop(ind)
class Edge:
def __init__(self, from_node, to_node, name):
self.from_node = from_node
self.to_node = to_node
self.name = name
def __str__(self):
return f'{self.from_node}->{self.to_node}'
class OperationDependencyGraph:
def __init__(self):
self.nodes = set()
self.edges = set()
self.from_to_dict = {}
self.feed_from_edges = {}
self.output_to_edges = {}
self.dependency_path = '../dependency.json'
self.threshold_length = 200
def add_edge(self, from_node, to_node, name):
if self.from_to_dict.__contains__((from_node, to_node)):
return self.from_to_dict[(from_node, to_node)]
edge = Edge(from_node, to_node, name)
feed_from_edges = self.feed_from_edges.get(to_node, [])
output_to_edges = self.output_to_edges.get(from_node, [])
feed_from_edges.append(edge)
output_to_edges.append(edge)
self.feed_from_edges[to_node] = feed_from_edges
self.output_to_edges[from_node] = output_to_edges
self.from_to_dict[(from_node, to_node)] = edge
self.nodes.add(from_node)
self.nodes.add(to_node)
self.edges.add(edge)
return edge
def remove_path_variable(self, pattern, target):
notations = pattern.findall(target)
for notation in notations:
target = target.replace(notation, "")
return target
def load_traffic_dependency(self, path):
with open(path, 'r', encoding="utf-8") as jsonfile:
data = jsonfile.read()
obj = json.loads(data)
return obj
def get_traffic_yaml_mapped_methods_map(self, dependency, yaml_method_map):
path_pattern = re.compile('\{(.*?)\}')
results = set()
for method in dependency:
nominal_name = self.remove_path_variable(path_pattern, method)
if nominal_name in yaml_method_map:
to_method = yaml_method_map[nominal_name]
feed_from_method_dependency = dependency[method]
for feed_from_method in feed_from_method_dependency:
nominal_feed_from_name = self.remove_path_variable(path_pattern, feed_from_method)
if nominal_feed_from_name in yaml_method_map:
from_method = yaml_method_map[nominal_feed_from_name]
to_method.dependency_from_traffic[from_method] = feed_from_method_dependency[feed_from_method]
from_method.output_to_method.add(to_method)
to_method.feed_from_method.add(from_method)
results.add((from_method, to_method))
return results
def get_traffic_map_with_yaml(self, path):
dependency = self.load_traffic_dependency(path)
yaml_method_map = self.get_yaml_method_path_map()
results = self.get_traffic_yaml_mapped_methods_map(dependency, yaml_method_map)
print(len(results), results)
def get_yaml_method_path_map(self):
path_pattern = re.compile('\{(.*?)\}')
result = {}
for method in self.nodes:
result[str(method.method_type).upper() + "-" + self.remove_path_variable(path_pattern,
method.method_path)] = method
return result
def add_node(self, node):
return self.nodes.add(node)
def get_output_edges(self, from_node):
return self.output_to_edges.get(from_node, [])
def get_feed_from_edges(self, to_node):
return self.feed_from_edges.get(to_node, [])
def draw(self, path='graph.txt'):
data_graph = Digraph()
for node in self.nodes:
data_graph.node(f'{node.method_type.upper()} {node.method_path}')
for edge in self.edges:
from_node = edge.from_node
to_node = edge.to_node
data_graph.edge(f'{to_node.method_type.upper()} {to_node.method_path}',
f'{from_node.method_type.upper()} {from_node.method_path}', edge.name)
# data_graph.render(path)
with open(path, 'w+') as output_graph:
output_graph.writelines(data_graph.source)
def generate_graph_sequence(self, method):
paths = []
covered_apis = set()
def traverse_path_recursive(method, path):
if method in path:
paths.append(path.duplicate())
elif len(method.output_to_method) == 0:
tmp = path.duplicate()
tmp.append(method)
covered_apis.add(method)
paths.append(tmp)
else:
for child in sorted(method.output_to_method, key=lambda x: x.method_signature):
if method.method_path == child.method_path and child.crud < method.crud:
# avoid empty path
if len(path) > 0:
paths.append(path.duplicate())
continue
if len(paths) > self.threshold_length and len(path) > 0:
continue
path.append(method)
covered_apis.add(method)
traverse_path_recursive(child, path)
path.remove(method)
traverse_path_recursive(method, Path())
return paths, covered_apis
# FIXME: should remove, only for debugging
def print_path(self, methods):
method_signatures = []
for method in methods:
method_signatures.append(f'{method.method_type} {method.method_path}')
res = ' -> '.join(method_signatures)
print(res)
return res
def extend_sequence(self, path):
feed_from_method = None
current_method = path[0]
sequence = Sequence([])
for i, method in enumerate(path):
required_property_dict = {}
if i != 0:
feed_from_method = path[i - 1]
current_method = method
assert feed_from_method != None
# check for crud and traffic dependency
required_property_dict = current_method.required_feed_parameter.get(feed_from_method.method_name,
{})
sequence.add_method(current_method)
# we do not add reference for the first one
if i == 0:
continue
# add dependency in the yaml (forward analysis)
for request_parameter in required_property_dict.keys():
nominal_request_parameters = current_method.get_request_paramter_by_property_name(
[request_parameter])
nominal_feed_response_parameters = feed_from_method.get_response_paramter_by_property_name(
[required_property_dict[request_parameter]])
for request_parameter in nominal_request_parameters:
for response_parameter in nominal_feed_response_parameters:
sequence.add_def(i - 1, response_parameter)
sequence.add_ref(i, feed_from_method, response_parameter,
request_parameter)
# add dependency from traffic
if current_method.dependency_from_traffic.__contains__(feed_from_method):
for dependency in current_method.dependency_from_traffic[feed_from_method]:
request_array, response_array = dependency
for request_parameter in request_array:
for response_parameter in response_array:
sequence.add_def(i - 1, response_parameter)
sequence.add_ref(i, feed_from_method, response_parameter, request_parameter)
assert len(path) == len(sequence)
return sequence
def generate_sequence(self, simple=False, fast=False):
crud_sort_map = {
"head": 1,
"post": 2,
"get": 3,
"put": 4,
"patch": 5,
"delete": 6,
}
raw_sequences = []
wrapped_sequences = []
covered_apis = set()
all_methods = sorted(self.nodes, key=lambda item: item.method_signature)
for method in all_methods:
if simple:
raw_sequences.append([method])
continue
if fast and np.random.random() < 0.5:
continue
# FIXME: need review
if len(method.output_to_method) == 0 and len(method.feed_from_method) > 0 and method in covered_apis:
continue
if method in covered_apis and method.crud > crud_sort_map['post']:
continue
print('traversing path', method)
paths, covered = self.generate_graph_sequence(method)
covered_apis = covered_apis.union(covered)
raw_sequences.extend(paths)
for seq in raw_sequences:
self.print_path(seq)
sequence = self.extend_sequence(seq)
sequence.origin = SequenceOrigin.ODG
wrapped_sequences.append(sequence)
print('hashing')
result = set(wrapped_sequences)
print("Generate sequences # : ", len(raw_sequences), ' extending sequences # : ', len(wrapped_sequences),
' result ', len(result))
return result
def get_single_node_sequence(self):
result = []
for method in self.nodes:
result.append(Sequence([method]))
return result