Skip to content

Commit

Permalink
Merge pull request #128 from Yanbo0101/estimate
Browse files Browse the repository at this point in the history 
fix estimate bug
  • Loading branch information
@sunya-ch
sunya-ch authored Aug 28, 2023
2 parents 0bcf63a + d882509 commit b4c9d3d
Showing 1 changed file with 69 additions and 65 deletions.
134 changes: 69 additions & 65 deletions cmd/main.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -348,6 +348,7 @@ def estimate(args):
from estimate import load_model, default_predicted_col_func, compute_error

inputs = args.input.split(",")
energy_sources = args.energy_source.split(",")
input_query_results_list = []
for input in inputs:
response = load_json(data_path, input)
Expand All @@ -372,71 +373,74 @@ def estimate(args):
print("invalid output type. please use AbsPower or DynPower", args.output_type)
exit()

if args.pipeline_name:
pipeline_names = [args.pipeline_name]
else:
pipeline_names = list_pipelines(data_path, args.energy_source, args.output_type)
energy_components = PowerSourceMap[args.energy_source]
best_result = None
best_model_path = None
best_mae = None
print("Pipelines: ", pipeline_names)
for pipeline_name in pipeline_names:
pipeline_path = get_pipeline_path(data_path, pipeline_name=pipeline_name)
pipeline_metadata = load_metadata(pipeline_path)
if pipeline_metadata is None:
print("no metadata for pipeline {}.".format(pipeline_name))
continue
pipeline = get_pipeline(pipeline_name, args.profile, pipeline_metadata["isolator"], pipeline_metadata["abs_trainers"], pipeline_metadata["dyn_trainers"], args.energy_source, valid_fg)
if pipeline is None:
print("cannot get pipeline {}.".format(pipeline_name))
continue
for fg in valid_fg:
print(" Feature Group: ", fg)
abs_data, dyn_data, power_labels = pipeline.prepare_data_from_input_list(input_query_results_list, energy_components, args.energy_source, fg.name)
group_path = get_model_group_path(data_path, ot, fg, args.energy_source, assure=False, pipeline_name=pipeline_name)
model_names = list_model_names(group_path)
if args.model_name:
if args.model_name not in model_names:
print("model: {} is not availble in pipeline {}, continue. available models are {}".format(args.model_name, pipeline_name, model_names))
continue
model_names = [args.model_name]
for model_name in model_names:
model_path = os.path.join(group_path, model_name)
model = load_model(model_path)
if args.output_type == ModelOutputType.AbsPower:
data = abs_data
else:
data = dyn_data
predicted_power_map, data_with_prediction = model.append_prediction(data)
max_mae = None
for energy_component, _ in predicted_power_map.items():
label_power_columns = [col for col in power_labels if energy_component in col]
predicted_power_colname = default_predicted_col_func(energy_component)
sum_power_label = data.groupby([TIMESTAMP_COL]).mean()[label_power_columns].sum(axis=1).sort_index()
sum_predicted_power = data_with_prediction.groupby([TIMESTAMP_COL]).sum().sort_index()[predicted_power_colname]
mae, mse = compute_error(sum_power_label, sum_predicted_power)
if max_mae is None or mae > max_mae:
max_mae = mae
if best_mae is None or max_mae < best_mae:
# update best
best_result = data_with_prediction.copy()
best_model_path = model_path
best_mae = max_mae
print(" Model {}: ".format(model_name), max_mae)

# save best result
if best_model_path is not None:
print("Energy consumption is predicted by {}".format(best_model_path.replace(data_path, "")))
print("MAE = ", best_mae)
output_folder = os.path.join(data_path, args.output)
if not os.path.exists(output_folder):
os.mkdir(output_folder)
# save model
import shutil
shutil.make_archive(os.path.join(output_folder, "model"), 'zip', best_model_path)
# save result
save_csv(output_folder, "estimation_result", best_result)
for energy_source in energy_sources:
if args.pipeline_name:
pipeline_names = [args.pipeline_name]
else:
pipeline_names = list_pipelines(data_path, energy_source, args.output_type)
energy_components = PowerSourceMap[energy_source]
best_result = None
best_model_path = None
best_mae = None
print("Pipelines: ", pipeline_names)
for pipeline_name in pipeline_names:
pipeline_path = get_pipeline_path(data_path, pipeline_name=pipeline_name)
pipeline_metadata = load_metadata(pipeline_path)
if pipeline_metadata is None:
print("no metadata for pipeline {}.".format(pipeline_name))
continue
pipeline = get_pipeline(pipeline_name, args.extractor, args.profile, pipeline_metadata["isolator"], pipeline_metadata["abs_trainers"], pipeline_metadata["dyn_trainers"], energy_sources, valid_fg)
if pipeline is None:
print("cannot get pipeline {}.".format(pipeline_name))
continue
for fg in valid_fg:
print(" Feature Group: ", fg)
abs_data, dyn_data, power_labels = pipeline.prepare_data_from_input_list(input_query_results_list, energy_components, energy_source, fg.name)
group_path = get_model_group_path(data_path, ot, fg, energy_source, assure=False, pipeline_name=pipeline_name)
model_names = list_model_names(group_path)
if args.model_name:
if args.model_name not in model_names:
print("model: {} is not availble in pipeline {}, continue. available models are {}".format(args.model_name, pipeline_name, model_names))
continue
model_names = [args.model_name]
for model_name in model_names:
model_path = os.path.join(group_path, model_name)
model = load_model(model_path)
if args.output_type == ModelOutputType.AbsPower:
data = abs_data
else:
data = dyn_data
predicted_power_map, data_with_prediction = model.append_prediction(data)
max_mae = None
for energy_component, _ in predicted_power_map.items():
label_power_columns = [col for col in power_labels if energy_component in col]
predicted_power_colname = default_predicted_col_func(energy_component)
sum_power_label = data.groupby([TIMESTAMP_COL]).mean()[label_power_columns].sum(axis=1).sort_index()
sum_predicted_power = data_with_prediction.groupby([TIMESTAMP_COL]).sum().sort_index()[predicted_power_colname]
mae, mse = compute_error(sum_power_label, sum_predicted_power)
if max_mae is None or mae > max_mae:
max_mae = mae
if best_mae is None or max_mae < best_mae:
# update best
best_result = data_with_prediction.copy()
best_model_path = model_path
best_mae = max_mae
print(" Model {}: ".format(model_name), max_mae)

# save best result
if best_model_path is not None:
print("Energy consumption of energy source {} is predicted by {}".format(energy_source, best_model_path.replace(data_path, "")))
print("MAE = ", best_mae)
output_folder = os.path.join(data_path, args.output)
if not os.path.exists(output_folder):
os.mkdir(output_folder)
# save model
import shutil
best_model = "{}_model".format(energy_source)
shutil.make_archive(os.path.join(output_folder, best_model), 'zip', best_model_path)
# save result
estimation_result = "{}_estimation_result".format(energy_source)
save_csv(output_folder, estimation_result, best_result)

if __name__ == "__main__":
# Create an ArgumentParser object
Expand Down

0 comments on commit b4c9d3d

Please sign in to comment.