Skip to content

Commit

Permalink
Refactoring
Browse files Browse the repository at this point in the history
  • Loading branch information
Juliette-Gerbaux committed Jul 30, 2024
1 parent 627aeb2 commit a5dcb1f
Show file tree
Hide file tree
Showing 7 changed files with 123 additions and 129 deletions.
44 changes: 13 additions & 31 deletions src/andromede/thermal_heuristic/model.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -45,24 +45,6 @@ class ModelEditor:
def __init__(self, initial_model: Model) -> None:
self.initial_model = initial_model

def keep_id(self) -> str:
return self.initial_model.id

def keep_parameters(self) -> List[Parameter]:
return [p for p in self.initial_model.parameters.values()]

def keep_ports(self) -> List[ModelPort]:
return [p for p in self.initial_model.ports.values()]

def keep_port_fields_definitions(self) -> List[PortFieldDefinition]:
return [p for p in self.initial_model.port_fields_definitions.values()]

def keep_constraints(self) -> List[Constraint]:
return [c for c in self.initial_model.constraints.values()]

def keep_objective_operational_contribution(self) -> Optional[ExpressionNode]:
return self.initial_model.objective_operational_contribution

def linearize_variables(self) -> List[Variable]:
return [
float_variable(
Expand Down Expand Up @@ -140,13 +122,13 @@ class AccurateModelBuilder(ModelEditor):
def __init__(self, initial_model: Model) -> None:
super().__init__(initial_model)
THERMAL_CLUSTER_MODEL_LP = model(
id=self.keep_id(),
parameters=self.keep_parameters(),
id=self.initial_model.id,
parameters=self.initial_model.parameters.values(),
variables=self.linearize_variables(),
ports=self.keep_ports(),
port_fields_definitions=self.keep_port_fields_definitions(),
constraints=self.keep_constraints(),
objective_operational_contribution=self.keep_objective_operational_contribution(),
ports=self.initial_model.ports.values(),
port_fields_definitions=self.initial_model.port_fields_definitions.values(),
constraints=self.initial_model.constraints.values(),
objective_operational_contribution=self.initial_model.objective_operational_contribution,
)
self.model = THERMAL_CLUSTER_MODEL_LP

Expand All @@ -157,13 +139,13 @@ def __init__(self, initial_model: Model) -> None:
integer_variables = self.get_name_integer_variables()

THERMAL_CLUSTER_MODEL_FAST = model(
id=self.keep_id(),
parameters=self.keep_parameters(),
id=self.initial_model.id,
parameters=self.initial_model.parameters.values(),
variables=self.fix_integer_variables_to_zero_and_keep_others(),
ports=self.keep_ports(),
port_fields_definitions=self.keep_port_fields_definitions(),
ports=self.initial_model.ports.values(),
port_fields_definitions=self.initial_model.port_fields_definitions.values(),
constraints=self.filter_constraints_on_variable(integer_variables),
objective_operational_contribution=self.keep_objective_operational_contribution(),
objective_operational_contribution=self.initial_model.objective_operational_contribution,
)

self.model = THERMAL_CLUSTER_MODEL_FAST
Expand All @@ -175,8 +157,8 @@ def __init__(self, initial_model: Model) -> None:
generation_variable = ["generation"]

THERMAL_CLUSTER_MODEL_ACCURATE_HEURISTIC = model(
id=self.keep_id(),
parameters=self.keep_parameters(),
id=self.initial_model.id,
parameters=self.initial_model.parameters.values(),
variables=self.filter_and_linearize_variables(generation_variable),
constraints=self.filter_constraints_on_variable(generation_variable),
objective_operational_contribution=(var("nb_on")).sum().expec(),
Expand Down
126 changes: 63 additions & 63 deletions src/andromede/thermal_heuristic/problem.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -76,23 +76,24 @@ def __init__(
port_types: List[PortType],
models: List[Model],
number_week: int,
list_scenario: List[int],
number_scenario: int,
) -> None:
lib = library_thermal_problem(
port_types, models, lp_relaxation, fast, initial_thermal_model
)
self.number_week = number_week
self.list_scenario = list_scenario
self.number_scenario = number_scenario
self.network = get_network(data_dir / "components.yml", lib)
self.number_hours = number_hours
self.initial_thermal_model = initial_thermal_model
self.database = self.get_database(
data_dir,
"components.yml",
get_cluster_id(self.network, initial_thermal_model.id),
number_hours,
fast,
timesteps=number_hours * number_week,
scenarios=number_scenario,
)
self.initial_thermal_model = initial_thermal_model

def get_main_resolution_step(self, week: int, scenario: int) -> ResolutionStep:
main_resolution_step = ResolutionStep(
Expand All @@ -114,14 +115,12 @@ def update_database_accurate(
list_cluster_id: Optional[list[str]],
) -> None:
if list_cluster_id is None:
list_cluster_id = get_cluster_id(
self.network, self.initial_thermal_model.id
)
list_cluster_id = self.get_milp_heuristic_components()
for cluster in list_cluster_id:
self.database.convert_to_time_scenario_series_data(
ComponentParameterIndex(cluster, "nb_units_min"),
self.number_hours * self.number_week,
len(self.list_scenario),
self.number_scenario,
)
nb_on = output.component(cluster).var("nb_on").value[0] # type:ignore

Expand All @@ -138,7 +137,7 @@ def update_database_accurate(
def update_database_fast_before_heuristic(
self, output: OutputValues, week: int, scenario: int
) -> None:
for cluster in get_cluster_id(self.network, self.initial_thermal_model.id):
for cluster in self.get_milp_heuristic_components():
pmax = self.database.get_value(
ComponentParameterIndex(cluster, "p_max"), 0, 0
)
Expand All @@ -148,7 +147,7 @@ def update_database_fast_before_heuristic(
self.database.convert_to_time_scenario_series_data(
ComponentParameterIndex(cluster, "n_guide"),
self.number_hours * self.number_week,
len(self.list_scenario),
self.number_scenario,
)

for i, t in enumerate(
Expand All @@ -174,9 +173,7 @@ def update_database_fast_after_heuristic(
list_cluster_id: Optional[list[str]],
) -> None:
if list_cluster_id is None:
list_cluster_id = get_cluster_id(
self.network, self.initial_thermal_model.id
)
list_cluster_id = self.get_milp_heuristic_components()
for cluster in list_cluster_id:
pmin = self.database.get_value(
ComponentParameterIndex(cluster, "p_min"), 0, 0
Expand All @@ -191,7 +188,7 @@ def update_database_fast_after_heuristic(
self.database.convert_to_time_scenario_series_data(
ComponentParameterIndex(cluster, "min_generating"),
self.number_hours * self.number_week,
len(self.list_scenario),
self.number_scenario,
)

min_gen = np.minimum(
Expand Down Expand Up @@ -340,39 +337,46 @@ def get_database(
self,
data_dir: Path,
yml_file: str,
cluster_id: List[str],
hours_in_week: int,
fast: bool,
timesteps: int,
scenarios: int,
) -> DataBase:
components_file = get_input_components(data_dir / yml_file)
database = build_data_base(components_file, data_dir)

complete_database_with_cluster_parameters(
database,
list_cluster_id=cluster_id,
dir_path=data_dir,
list_cluster_id=self.get_milp_heuristic_components(),
hours_in_week=hours_in_week,
timesteps=timesteps,
scenarios=scenarios,
)

if fast:
self.complete_database_for_fast_heuristic(
database, cluster_id, hours_in_week
database, self.get_milp_heuristic_components(), hours_in_week
)

return database

def get_milp_heuristic_components(self) -> list[str]:
return [
c.id
for c in self.network.components
if c.model.id == self.initial_thermal_model.id
]


def complete_database_with_cluster_parameters(
database: DataBase, list_cluster_id: List[str], dir_path: Path, hours_in_week: int
database: DataBase,
list_cluster_id: List[str],
hours_in_week: int,
timesteps: int,
scenarios: int,
) -> None:
for cluster_id in list_cluster_id:
data = get_data(
dir_path,
"components.yml",
cluster_id,
)

if type(data["max_generating"]) is float:
if type(database.get_data(cluster_id, "max_generating")) is ConstantData:
database.add_data(
cluster_id,
"max_failure",
Expand All @@ -381,15 +385,17 @@ def complete_database_with_cluster_parameters(
database.add_data(
cluster_id,
"nb_units_max_min_down_time",
ConstantData(data["nb_units_max"]),
database.get_data(cluster_id, "nb_units_max"),
)

else:
(
max_units,
max_failures,
nb_units_max_min_down_time,
) = compute_cluster_parameters(hours_in_week, data)
) = compute_cluster_parameters(
hours_in_week, database, cluster_id, timesteps, scenarios
)
database.add_data(
cluster_id,
"nb_units_max",
Expand All @@ -408,28 +414,43 @@ def complete_database_with_cluster_parameters(


def compute_cluster_parameters(
hours_in_week: int, data: Dict
hours_in_week: int,
database: DataBase,
cluster_id: str,
timesteps: int,
scenarios: int,
) -> tuple[pd.DataFrame, pd.DataFrame, pd.DataFrame]:
max_generating_shape = data["max_generating"].shape
if len(max_generating_shape) == 1:
max_generating = pd.DataFrame(
data["max_generating"].reshape((max_generating_shape[0], 1))
)
else:
max_generating = pd.DataFrame(data["max_generating"])
max_units = get_max_unit(data["p_max"], data["nb_units_max"], max_generating)
database.convert_to_time_scenario_series_data(
ComponentParameterIndex(cluster_id, "max_generating"),
timesteps=timesteps,
scenarios=scenarios,
)
max_units = get_max_unit(
database.get_value(ComponentParameterIndex(cluster_id, "p_max"), 0, 0),
database.get_value(ComponentParameterIndex(cluster_id, "nb_units_max"), 0, 0),
database.get_data(
cluster_id, "max_generating"
).time_scenario_series, # type:ignore
)
max_failures = get_max_failures(max_units, hours_in_week)
nb_units_max_min_down_time = get_max_unit_for_min_down_time(
int(max(data["d_min_up"], data["d_min_down"])), max_units, hours_in_week
int(
max(
database.get_value(
ComponentParameterIndex(cluster_id, "d_min_up"), 0, 0
),
database.get_value(
ComponentParameterIndex(cluster_id, "d_min_down"), 0, 0
),
)
),
max_units,
hours_in_week,
)

return max_units, max_failures, nb_units_max_min_down_time


def get_cluster_id(network: Network, cluster_model_id: str) -> list[str]:
return [c.id for c in network.components if c.model.id == cluster_model_id]


def get_network(compo_file: Path, lib: Library) -> Network:
components_file = get_input_components(compo_file)
components_input = resolve_components_and_cnx(components_file, lib)
Expand All @@ -453,24 +474,3 @@ def edit_thermal_model(
else:
thermal_model = initial_thermal_model
return thermal_model


def get_data(
data_dir: Path,
yml_file: str,
id_cluster: str,
) -> Dict:
compo_file = data_dir / yml_file

with compo_file.open() as c:
components_file = parse_yaml_components(c)
cluster = [c for c in components_file.components if c.id == id_cluster][0]
parameters = {
p.name: (
p.value
if p.value is not None
else np.loadtxt(data_dir / (p.timeseries + ".txt")) # type:ignore
)
for p in cluster.parameters # type:ignore
}
return parameters # type:ignore
4 changes: 2 additions & 2 deletions tests/functional/test_thermal_heuristic_fast_min_down_not_respected.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -41,10 +41,10 @@ def test_fast_heuristic() -> None:
port_types=[],
models=[],
number_week=1,
list_scenario=list(range(1)),
number_scenario=1,
)

cluster = "G1"
cluster = thermal_problem_builder.get_milp_heuristic_components()[0]

pmax = thermal_problem_builder.database.get_value(
ComponentParameterIndex(cluster, "p_max"), 0, 0
Expand Down
Loading

0 comments on commit a5dcb1f

Please sign in to comment.