def test_log():
with open(os.path.join(os.path.dirname(__file__), "log.txt")) as f:
flag = False
for row in f:
if row[:5] == "train":
file_kbn = list(row.split())
elif row[:4] == "eval":
file_kbn = list(row.split())
elif row[:6] == "Failed":
flag = False
elif row[:6] == "Solved":
flag = True
# print(row, file_kbn)
if flag and row[0] == "[":
op_list = eval(row.strip())
p = Runner(int(file_kbn[1]), file_kbn[0], verbose=True)
for op in op_list:
p.run(op)
ac, wa = p.eval_test(verbose=True)
if wa == 0 and file_kbn[0] == "eval":
# print(file_kbn[1], op_list)
pass
if wa > 0:
print(op_list)
print(ac, wa)
assert p.eval_distance() == 0
def main(dir_path, time_limit):
output_id_list = []
output_list = []
for i, file_name in enumerate(list(sorted(os.listdir(dir_path)))):
if file_name[-5:] == ".json":
p = Runner()
with open(dir_path + file_name, "r") as file:
data = json.load(file)
p.initialize_from_data(data)
try:
p.auto_run(time_limit=time_limit)
solved_dict = p.output()
for j in solved_dict.keys():
assert len(solved_dict[j]) <= 3
output_id_list.append(f'{file_name[:-5]}_{j}')
output_list.append(" ".join(solved_dict[j]))
except:
for j in range(len(data["test"])):
output_id_list.append(f'{file_name[:-5]}_{j}')
output_list.append("")
res_df = pd.DataFrame({"output_id": output_id_list, "output": output_list})
# print(res_df)
res_df.to_csv("submission.csv", index=False)
def test_488():
p = Runner(88, "eval", verbose=True)
p.run("connect")
p.run("count_hole")
p.run("freq")
p.run("sort_ascending")
p.run("trim_background")
def main():
runner = None
try:
config = Config.load()
init_logging(config)
runner = Runner()
runner.open(config)
runner.run()
return 0
except KeyboardInterrupt:
# if runner is not None:
# runner.close()
return 0
except Exception as ex:
_logger.exception(ex)
# no runner.close() to signal abnomal termination!
return 1
finally:
if runner is not None:
runner.close()
def test_58():
p = Runner(58, verbose=True)
p.run("mesh_split")
p.run("n_cell")
p.run("keep_max")
p.run("max_color")
p.run("paste_color_full")
def test_train_pipeline(fix_seed, config, gpus):
config = OmegaConf.create(config)
train_dataloader, test_dataloader = get_data_loaders(config=config)
lr_logger = LearningRateLogger()
model = build_model(model_conf=config.model)
runner = Runner(model=model, config=config.runner)
trainer = Trainer(
distributed_backend=config.runner.trainer.distributed_backend,
fast_dev_run=True,
gpus=gpus,
amp_level="O2",
row_log_interval=10,
callbacks=[lr_logger],
max_epochs=1,
weights_summary="top",
reload_dataloaders_every_epoch=False,
resume_from_checkpoint=None,
benchmark=False,
deterministic=True,
num_sanity_val_steps=5,
overfit_batches=0.0,
precision=32,
profiler=True,
)
trainer.fit(model=runner,
train_dataloader=train_dataloader,
val_dataloaders=test_dataloader)
def run_local(kbn="training", time_limit=0.2, verbose=True):
total_ac = 0
total_wa = 0
res_list = []
p_list = []
ind_list = []
assert kbn == "training" or kbn == "evaluation"
for ind in range(400):
if kbn == "training":
p = Runner(ind, "train", verbose=verbose)
elif kbn == "evaluation":
p = Runner(ind, "eval", verbose=verbose)
else:
raise ValueError
p.auto_run(time_limit=time_limit)
solved_dict = p.output()
for j in solved_dict.keys():
p_list.append(ind)
ind_list.append(j)
assert len(solved_dict[j]) <= 3
answer_str = p.answer_list[j]
if answer_str in solved_dict[j]:
# print(f'AC: {i, j}')
total_ac += 1
res_list.append(1)
else:
# print(f'WA: {i, j}')
total_wa += 1
res_list.append(0)
pct = 1 - total_ac / (total_ac + total_wa)
print(f'{kbn} done, AC: {total_ac}, total: {total_ac + total_wa}, {pct}')
res_arr = np.concatenate([
np.array(res_list).reshape((-1, 1)),
np.array(p_list).reshape((-1, 1)),
np.array(ind_list).reshape((-1, 1))
],
axis=1)
np.save(f'../local_eval_log/{kbn}-{time_limit}-{pct}', res_arr)
return None
def test_440():
p = Runner(40, "eval", verbose=True)
p.run("connect")
def test_221():
p = Runner(221, verbose=True)
def test_261():
p = Runner(261, verbose=True)
def test_654():
p = Runner(254, "eval", verbose=True)
p.run("connect")
p.run("auto_keep")
def test_217():
p = Runner(217, verbose=True)
def test_555():
p = Runner(155, "eval", verbose=True)
p.run("trim_background")
def test_578():
p = Runner(178, "eval", verbose=True)
p.run("trim_background")
def test_586():
p = Runner(186, "eval", verbose=True)
p.run("trim_background")
def test_64():
p = Runner(64, verbose=True)
def test_607():
p = Runner(207, "eval", verbose=True)
p.run("mesh_split")
p.run("collect_max")
def test_61():
p = Runner(61, verbose=True)
def test_756():
p = Runner(356, "eval", verbose=True)
p.run("color")
def test_738():
p = Runner(338, "eval", verbose=True)
p.run("change_background")
p.run("connect4")
p.run("auto_paste_a")
def test_719():
p = Runner(319, "eval", verbose=True)
from src.runner.runner import Runner
from src.adityaork.tree import predict_from_json
if __name__ == "__main__":
p = Runner(14, file_list="train", verbose=True)
s = predict_from_json(p.task_json)
print(s)
p = Runner(15, file_list="train", verbose=True)
s = predict_from_json(p.task_json)
print(s)
def test_85():
p = Runner(85, verbose=True)
def test_551():
p = Runner(151, "eval", verbose=True)
def test_134():
p = Runner(134, verbose=True)
p.run("divide")
def test_561():
p = Runner(161, "eval", verbose=True)
p.run("trim_background")
def test_140():
p = Runner(140, verbose=True)
def test_124():
p = Runner(124, verbose=True)
p.run("change_background")
def test_652():
p = Runner(252, "eval", verbose=True)
p.run("color")
p.run("auto_keep")
p.run("align")