def use_exe_save_and_analysis_str_and_csv(
save_path: str, compared_csv_path: str, csv_name: str,
model_path: str, test_dir_path: str,
test_len: int, word_len: int, thresh_num: float
) -> None:
"""
用lstm生成的答案的csv文件来生成
"""
with open(save_path, "rb") as f:
exe_res = pickle.load(f)
predict_by_network = PredictByNetwork(model_path, test_dir_path,
test_len, word_len, thresh_num)
g = GenerateAns(csv_name, "sample_submission.csv")
compared_csv = GenerateAns("", compared_csv_path)
while True:
t = g.get()
compared_csv.get()
if t is None:
break
b1 = compare_test(exe_res[t[0]], exe_res[t[1]])
b2 = analysis_str(t[0], t[1])
b3 = ((judge(exe_res[t[0]][0], exe_res[t[1]][0]) and exe_res[t[0]][0].result != -1)
or (judge(exe_res[t[0]][1], exe_res[t[1]][1]) and exe_res[t[1]][0].result != -1))\
and compared_csv.get_cur_value() == 1
p = 1 if b1 or b2 or b3 else 0
if not b1 and not b2:
print("hehehe")
g.add(p)
def use_exe_save_and_analysis_str(save_path: str, csv_name: str) -> None:
"""
有点点效果，提升到了0.74
"""
with open(save_path, "rb") as f:
exe_res = pickle.load(f)
g = GenerateAns(csv_name, "sample_submission.csv")
while True:
t = g.get()
if t is None:
break
g.add(1 if compare_test(exe_res[t[0]], exe_res[t[1]]) or analysis_str(t[0], t[1]) else 0)
def use_exe_save_and_analysis_str_and_xgboost(save_path: str, csv_name: str,
test_index_map: str, network_result_path: str) -> None:
with open(save_path, "rb") as f:
exe_res = pickle.load(f)
gcr = GetCompareResult(test_index_map, network_result_path)
g = GenerateAns(csv_name, "sample_submission.csv")
while True:
t = g.get()
if t is None:
break
xgboost_bool = gcr.get(t[0], t[1])
# xgboost_bool = False
g.add(1 if compare_test(exe_res[t[0]], exe_res[t[1]]) or analysis_str(t[0], t[1]) or xgboost_bool else 0)
def use_exe_save_and_analysis_str_input_output_loop(save_path: str, csv_name: str) -> None:
"""
效果不佳，调出的最好成绩为0.70
"""
with open(save_path, "rb") as f:
exe_res = pickle.load(f)
g = GenerateAns(csv_name, "sample_submission.csv")
while True:
t = g.get()
if t is None:
break
g.add(1 if compare_test(exe_res[t[0]], exe_res[t[1]])
or analysis_str(t[0], t[1])
or analysis_input_output_loop_num(t[0], t[1])
else 0)
def use_exe_save_and_analysis_str_and_network(
save_path: str, csv_name: str,
model_path: str, test_dir_path: str,
test_len: int, word_len: int, thresh_num: float
) -> None:
"""
辅助以lstm
"""
with open(save_path, "rb") as f:
exe_res = pickle.load(f)
predict_by_network = PredictByNetwork(model_path, test_dir_path,
test_len, word_len, thresh_num)
g = GenerateAns(csv_name, "sample_submission.csv")
while True:
t = g.get()
if t is None:
break
g.add(1 if compare_test(exe_res[t[0]], exe_res[t[1]])
or analysis_str(t[0], t[1])
or predict_by_network.predict(t[0], t[1])
else 0)
def use_exe_save_and_analysis_str_and_xgboost_strict(save_path: str, csv_name: str,
test_index_map: str, network_result_path: str) -> None:
# 判断更为严格：只有在运行的时候有运行失败的情况才进行后面的补充判断
with open(save_path, "rb") as f:
exe_res = pickle.load(f)
gcr = GetCompareResult(test_index_map, network_result_path)
g = GenerateAns(csv_name, "./build/sample_submission.csv")
while True:
t = g.get()
if t is None:
break
xgboost_bool = gcr.get(t[0], t[1])
min_succ_num = min(sum([1 if x.result == 0 else 0 for x in exe_res[t[0]]]),
sum([1 if x.result == 0 else 0 for x in exe_res[t[1]]]))
# xgboost_bool = False
same_num = compare_test7(exe_res[t[0]], exe_res[t[1]])
g.add(1 if
same_num == len(exe_res[t[0]]) or analysis_str(t[0], t[1])
or ((min_succ_num == 0 or same_num >= 1) and xgboost_bool)
else 0)