def _build_req_dict_for_map(trace_link_candidates,
dataset: Dataset,
reverse_compare=False):
trace_link_candidates = list(
dict.fromkeys(trace_link_candidates
)) # Deterministic duplicate removal (set is not stable)
if len(trace_link_candidates) < dataset.num_original_links():
# trace_link_candidates does not contain all possible link between reqs and code
# -> Add missing links as dummy links with 0 similarity
code_filenames = dataset.all_original_code_file_names(True)
req_filenames = dataset.all_original_req_file_names(True)
for req_name in req_filenames:
for code_name in code_filenames:
no_similarity = 1 if reverse_compare else 0
dummy_trace_link = TraceLink(MockEmbedding(req_name),
MockEmbedding(code_name),
no_similarity)
if not dummy_trace_link in trace_link_candidates:
trace_link_candidates.append(dummy_trace_link)
assert len(trace_link_candidates) == dataset.num_original_links(
), f"{len(trace_link_candidates)} != {dataset.num_original_links()}"
req_dict = {
} # req_dict["req_name"] = [(sim_to_code_1: float, relevant: bool), (sim_to_code_2, relevant), ...]
sol_matrix_copy = Util.deep_copy(dataset.solution_matrix(
)) # Use copy to track false negatives and avoid duplicate trace links
for trace_link in trace_link_candidates:
req_name = trace_link.get_req_key(dataset.keys_with_extension())
code_name = trace_link.get_code_key(dataset.keys_with_extension())
sim_rel_tuple_to_add = (trace_link.similarity, False)
if sol_matrix_copy.contains_req_code_pair(req_name, code_name):
sim_rel_tuple_to_add = (trace_link.similarity, True)
sol_matrix_copy.remove_trace_pair(req_name, code_name)
if req_name in req_dict:
req_dict[req_name].append(sim_rel_tuple_to_add)
else:
req_dict[req_name] = [sim_rel_tuple_to_add]
if PRINT_FALSE_NEGATIVES:
_print_false_negatives(sol_matrix_copy)
return req_dict
def evaluateMAPRecall(trace_link_candidates,
dataset: Dataset,
reverse_compare=False):
"""
trace_link_candidates: List of TraceLink-objects
"""
if not trace_link_candidates:
text = "No Trace Link candidates!"
log.info(text)
return 0, 0
req_dict = _build_req_dict_for_map(trace_link_candidates, dataset,
reverse_compare)
recall_map_dict = {}
for k in range(1, len(dataset.all_original_code_file_names()) + 1):
map_k, rec_k = Util.calculate_mean_average_precision_and_recall(
req_dict, k, dataset.num_reqs(),
dataset.solution_matrix()._number_of_trace_links, reverse_compare)
recall_map_dict[rec_k] = map_k
print(f"{k}: {rec_k}, {map_k}")
return recall_map_dict
