Skip to content

KPI Availability Forecast Curtailment Events

KpiAvailabilityForecastCurtailmentEvents(perfdb)

Class used for handling power forecast curtailment events. Can be accessed via perfdb.kpis.availability.forecasts.curtailment_events.

Parameters:

  • perfdb

    (PerfDB) –

    Top level object carrying all functionality and the connection handler.

Source code in echo_postgres/kpi_availability_forecast_curtailment_events.py
Python
def __init__(self, perfdb: e_pg.PerfDB) -> None:
    """Class used for handling power forecast curtailment events. Can be accessed via `perfdb.kpis.availability.forecasts.curtailment_events`.

    Parameters
    ----------
    perfdb : PerfDB
        Top level object carrying all functionality and the connection handler.
    """
    super().__init__(perfdb)

    from .kpi_availability_forecast_curtailment_event_assets import KpiAvailabilityForecastCurtailmentEventAssets

    self.assets = KpiAvailabilityForecastCurtailmentEventAssets(perfdb)

delete(ids)

Deletes power forecast curtailment events by ID.

Note: linked assets (in power_forecast_curtailment_event_assets) are cascade-deleted.

Parameters:

  • ids

    (list[int]) –

    List of curtailment event IDs to delete.

Returns:

  • int –

    Number of rows deleted.

Source code in echo_postgres/kpi_availability_forecast_curtailment_events.py
Python
@validate_call
def delete(self, ids: list[int]) -> int:
    """Deletes power forecast curtailment events by ID.

    Note: linked assets (in power_forecast_curtailment_event_assets) are cascade-deleted.

    Parameters
    ----------
    ids : list[int]
        List of curtailment event IDs to delete.

    Returns
    -------
    int
        Number of rows deleted.
    """
    query = sql.SQL("DELETE FROM performance.power_forecast_curtailment_events WHERE id = ANY({ids})").format(
        ids=sql.Literal(ids),
    )
    self._perfdb.conn.execute(query)
    deleted = self._perfdb.conn.rowcount
    logger.debug(f"Deleted {deleted} curtailment event(s).")
    return deleted

get(ids=None, revision_names=None, period=None, filter_type='and', output_type='pl.DataFrame')

Gets power forecast curtailment events with full details.

The most useful keys/columns returned are:

  • id
  • revision_name
  • description
  • curtailment_percentage
  • power_fraction_available
  • start_time
  • end_time
  • duration_hours
  • duration_days

Parameters:

  • ids

    (list[int] | None, default: None ) –

    List of event IDs to filter. By default None.

  • revision_names

    (list[str] | None, default: None ) –

    List of revision names to filter. By default None.

  • period

    (DateTimeRange | None, default: None ) –

    Time range; returns events that overlap with this range. By default None.

  • filter_type

    (Literal['and', 'or'], default: 'and' ) –

    How to treat multiple filters. By default "and".

  • output_type

    (Literal['dict', 'DataFrame', 'pl.DataFrame'], default: 'pl.DataFrame' ) –

    Output type. By default "pl.DataFrame".

Returns:

  • dict[int, dict[str, Any]] –

    In case output_type is "dict", returns {id: {col: val, ...}, ...}.

  • DataFrame –

    In case output_type is "DataFrame", returns a pandas DataFrame indexed by id.

  • DataFrame –

    In case output_type is "pl.DataFrame", returns a Polars DataFrame.

Source code in echo_postgres/kpi_availability_forecast_curtailment_events.py
Python
@validate_call
def get(
    self,
    ids: list[int] | None = None,
    revision_names: list[str] | None = None,
    period: DateTimeRange | None = None,
    filter_type: Literal["and", "or"] = "and",
    output_type: Literal["dict", "DataFrame", "pl.DataFrame"] = "pl.DataFrame",
) -> dict[int, dict[str, Any]] | pd.DataFrame | pl.DataFrame:
    """Gets power forecast curtailment events with full details.

    The most useful keys/columns returned are:

    - id
    - revision_name
    - description
    - curtailment_percentage
    - power_fraction_available
    - start_time
    - end_time
    - duration_hours
    - duration_days

    Parameters
    ----------
    ids : list[int] | None, optional
        List of event IDs to filter. By default None.
    revision_names : list[str] | None, optional
        List of revision names to filter. By default None.
    period : DateTimeRange | None, optional
        Time range; returns events that overlap with this range. By default None.
    filter_type : Literal["and", "or"], optional
        How to treat multiple filters. By default "and".
    output_type : Literal["dict", "DataFrame", "pl.DataFrame"], optional
        Output type. By default "pl.DataFrame".

    Returns
    -------
    dict[int, dict[str, Any]]
        In case output_type is "dict", returns {id: {col: val, ...}, ...}.
    pd.DataFrame
        In case output_type is "DataFrame", returns a pandas DataFrame indexed by id.
    pl.DataFrame
        In case output_type is "pl.DataFrame", returns a Polars DataFrame.
    """
    where = self._check_get_args(ids, revision_names, period, filter_type)
    query = sql.SQL(
        "SELECT * FROM performance.v_power_forecast_curtailment_events {where} ORDER BY revision_name, start_time",
    ).format(where=where)
    df = self._perfdb.conn.read_to_polars(query, schema_overrides=self._cols_schema)
    return convert_output(df, output_type, index_col="id")

get_defaults(years=None, months=None, object_names=None, revision_names=None, filter_type='and', output_type='pl.DataFrame')

Gets the per-asset / per-month resolution of the curtailment events that are the default for each (asset, year, month) according to availability_forecast_revision_assignments.

Backed by performance.v_power_forecast_defaults_curtailment_events. Emits one row per (asset, year, month, event) where the asset's revision assignment for that month points to the event's revision and the event's [start_time, end_time] overlaps the calendar month. curtailment_percentage is a fraction (0..1); month_impact is the calendar-month-relative fraction (0..1) weighted by curtailment_percentage.

Parameters:

  • years

    (list[int] | None, default: None ) –

    List of years to filter. By default None.

  • months

    (list[int] | None, default: None ) –

    List of months (1-12) to filter. By default None.

  • object_names

    (list[str] | None, default: None ) –

    List of object names to filter. By default None.

  • revision_names

    (list[str] | None, default: None ) –

    List of revision names to filter. By default None.

  • filter_type

    (Literal['and', 'or'], default: 'and' ) –

    How to treat multiple filters. By default "and".

  • output_type

    (Literal['dict', 'DataFrame', 'pl.DataFrame'], default: 'pl.DataFrame' ) –

    Output type. By default "pl.DataFrame".

Returns:

  • dict[Any, dict[str, Any]] –

    In case output_type is "dict", returns {(object_name, year, month, record_id): {col: val, ...}, ...}.

  • DataFrame –

    In case output_type is "DataFrame", returns a pandas DataFrame with MultiIndex (object_name, year, month, record_id).

  • DataFrame –

    In case output_type is "pl.DataFrame", returns a Polars DataFrame.

Source code in echo_postgres/kpi_availability_forecast_curtailment_events.py
Python
@validate_call
def get_defaults(
    self,
    years: list[int] | None = None,
    months: list[int] | None = None,
    object_names: list[str] | None = None,
    revision_names: list[str] | None = None,
    filter_type: Literal["and", "or"] = "and",
    output_type: Literal["dict", "DataFrame", "pl.DataFrame"] = "pl.DataFrame",
) -> dict[Any, dict[str, Any]] | pd.DataFrame | pl.DataFrame:
    """Gets the per-asset / per-month resolution of the curtailment events that are the default for each
    (asset, year, month) according to availability_forecast_revision_assignments.

    Backed by ``performance.v_power_forecast_defaults_curtailment_events``. Emits one row per
    (asset, year, month, event) where the asset's revision assignment for that month points to the
    event's revision and the event's [start_time, end_time] overlaps the calendar month.
    ``curtailment_percentage`` is a fraction (0..1); ``month_impact`` is the calendar-month-relative
    fraction (0..1) weighted by curtailment_percentage.

    Parameters
    ----------
    years : list[int] | None, optional
        List of years to filter. By default None.
    months : list[int] | None, optional
        List of months (1-12) to filter. By default None.
    object_names : list[str] | None, optional
        List of object names to filter. By default None.
    revision_names : list[str] | None, optional
        List of revision names to filter. By default None.
    filter_type : Literal["and", "or"], optional
        How to treat multiple filters. By default "and".
    output_type : Literal["dict", "DataFrame", "pl.DataFrame"], optional
        Output type. By default "pl.DataFrame".

    Returns
    -------
    dict[Any, dict[str, Any]]
        In case output_type is "dict", returns {(object_name, year, month, record_id): {col: val, ...}, ...}.
    pd.DataFrame
        In case output_type is "DataFrame", returns a pandas DataFrame with MultiIndex
        (object_name, year, month, record_id).
    pl.DataFrame
        In case output_type is "pl.DataFrame", returns a Polars DataFrame.
    """
    where = (
        WhereClauseBuilder(filter_type=filter_type)
        .add_any("year", years)
        .add_any("month", months)
        .add_any("object_name", object_names)
        .add_any("revision_name", revision_names)
        .build()
    )
    query = sql.SQL(
        "SELECT * FROM performance.v_power_forecast_defaults_curtailment_events {where} "
        "ORDER BY object_name, year, month, record_id",
    ).format(where=where)
    df = self._perfdb.conn.read_to_polars(query, schema_overrides=self._defaults_cols_schema)
    return convert_output(df, output_type, index_col=["object_name", "year", "month", "record_id"])

get_ids(revision_names=None, period=None, filter_type='and')

Gets IDs of power forecast curtailment events.

Parameters:

  • revision_names

    (list[str] | None, default: None ) –

    List of revision names to filter. By default None.

  • period

    (DateTimeRange | None, default: None ) –

    Time range; returns events that overlap with this range. By default None.

  • filter_type

    (Literal['and', 'or'], default: 'and' ) –

    How to treat multiple filters. By default "and".

Returns:

  • list[int] –

    List of curtailment event IDs.

Source code in echo_postgres/kpi_availability_forecast_curtailment_events.py
Python
@validate_call
def get_ids(
    self,
    revision_names: list[str] | None = None,
    period: DateTimeRange | None = None,
    filter_type: Literal["and", "or"] = "and",
) -> list[int]:
    """Gets IDs of power forecast curtailment events.

    Parameters
    ----------
    revision_names : list[str] | None, optional
        List of revision names to filter. By default None.
    period : DateTimeRange | None, optional
        Time range; returns events that overlap with this range. By default None.
    filter_type : Literal["and", "or"], optional
        How to treat multiple filters. By default "and".

    Returns
    -------
    list[int]
        List of curtailment event IDs.
    """
    where = self._check_get_args(None, revision_names, period, filter_type)
    query = sql.SQL("SELECT id FROM performance.v_power_forecast_curtailment_events {where} ORDER BY id").format(where=where)
    df = self._perfdb.conn.read_to_polars(query, schema_overrides=self._cols_schema)
    return df["id"].to_list()

insert(revision_name=None, description=None, curtailment_percentage=None, start_time=None, end_time=None, user_name=None, data_df=None, on_conflict='raise')

Inserts one or more power forecast curtailment events.

You can pass individual values to insert a single event, or a DataFrame for batch insert.

Parameters:

  • revision_name

    (str | None, default: None ) –

    Name of the forecast revision this event belongs to. Required for single insert. By default None.

  • description

    (str | None, default: None ) –

    Human-readable explanation of the curtailment. Required for single insert. By default None.

  • curtailment_percentage

    (float | None, default: None ) –

    Curtailment level as a percentage (must be in range (0, 90]). Required for single insert. By default None.

  • start_time

    (datetime | None, default: None ) –

    Planned start timestamp. Required for single insert. By default None.

  • end_time

    (datetime | None, default: None ) –

    Planned end timestamp (must be after start_time). Required for single insert. By default None.

  • user_name

    (str | None, default: None ) –

    Name of the user creating the event. Required for single insert. By default None.

  • data_df

    (DataFrame | None, default: None ) –

    DataFrame for batch insert. Required columns: revision_name, description, curtailment_percentage, start_time, end_time, user_name. When provided, individual parameters are ignored. By default None.

  • on_conflict

    (Literal['raise', 'ignore'], default: 'raise' ) –

    Behavior on unexpected conflict. By default "raise".

Returns:

  • int | list[int] | None –

    ID(s) of the inserted event(s).

Source code in echo_postgres/kpi_availability_forecast_curtailment_events.py
Python
@validate_call
def insert(
    self,
    revision_name: str | None = None,
    description: str | None = None,
    curtailment_percentage: float | None = None,
    start_time: datetime | None = None,
    end_time: datetime | None = None,
    user_name: str | None = None,
    data_df: pl.DataFrame | None = None,
    on_conflict: Literal["raise", "ignore"] = "raise",
) -> int | list[int] | None:
    """Inserts one or more power forecast curtailment events.

    You can pass individual values to insert a single event, or a DataFrame for batch insert.

    Parameters
    ----------
    revision_name : str | None, optional
        Name of the forecast revision this event belongs to. Required for single insert. By default None.
    description : str | None, optional
        Human-readable explanation of the curtailment. Required for single insert. By default None.
    curtailment_percentage : float | None, optional
        Curtailment level as a percentage (must be in range (0, 90]). Required for single insert. By default None.
    start_time : datetime | None, optional
        Planned start timestamp. Required for single insert. By default None.
    end_time : datetime | None, optional
        Planned end timestamp (must be after start_time). Required for single insert. By default None.
    user_name : str | None, optional
        Name of the user creating the event. Required for single insert. By default None.
    data_df : pl.DataFrame | None, optional
        DataFrame for batch insert. Required columns: revision_name, description,
        curtailment_percentage, start_time, end_time, user_name. When provided, individual
        parameters are ignored. By default None.
    on_conflict : Literal["raise", "ignore"], optional
        Behavior on unexpected conflict. By default "raise".

    Returns
    -------
    int | list[int] | None
        ID(s) of the inserted event(s).
    """
    df_schema = {
        "revision_name": pl.Utf8,
        "description": pl.Utf8,
        "curtailment_percentage": pl.Float64,
        "start_time": pl.Datetime("ms"),
        "end_time": pl.Datetime("ms"),
        "user_name": pl.Utf8,
    }

    if data_df is None:
        single_insert = True
        data_df = pl.DataFrame(
            {
                "revision_name": [revision_name],
                "description": [description],
                "curtailment_percentage": [curtailment_percentage],
                "start_time": [start_time],
                "end_time": [end_time],
                "user_name": [user_name],
            },
            schema=df_schema,
        )
    else:
        single_insert = False
        data_df = data_df.cast({c: t for c, t in df_schema.items() if c in data_df.columns})

    required_cols = ["revision_name", "description", "curtailment_percentage", "start_time", "end_time", "user_name"]
    for col in required_cols:
        if col not in data_df.columns:
            raise ValueError(f"data_df is missing required column '{col}'.")
        if data_df[col].is_null().any():
            raise ValueError(f"Column '{col}' contains null values but is required.")

    # validate curtailment_percentage range
    invalid_pct = data_df.filter((pl.col("curtailment_percentage") <= 0) | (pl.col("curtailment_percentage") > 90))
    if len(invalid_pct) > 0:
        raise ValueError("'curtailment_percentage' must be in range (0, 90].")

    # resolve revision_name → revision_id
    rev_names = data_df["revision_name"].unique().to_list()
    rev_ids = self._perfdb.kpis.availability.forecasts.revisions.get_ids(names=rev_names)
    if missing := set(rev_names) - set(rev_ids.keys()):
        raise ValueError(f"Revisions not found: {missing}")
    data_df = data_df.with_columns(pl.col("revision_name").replace_strict(rev_ids, return_dtype=pl.Int64).alias("revision_id"))
    data_df = data_df.drop("revision_name")

    # resolve user_name → user_id
    user_names = data_df["user_name"].unique().to_list()
    user_ids = self._perfdb.users.instances.get_ids(names=user_names)
    if missing := set(user_names) - set(user_ids.keys()):
        raise ValueError(f"Users not found: {missing}")
    data_df = data_df.with_columns(pl.col("user_name").replace_strict(user_ids, return_dtype=pl.Int64).alias("user_id"))
    data_df = data_df.drop("user_name")

    if_exists = "append" if on_conflict == "ignore" else "skip_row_check"

    ids_df = self._perfdb.conn.polars_to_sql(
        df=data_df,
        table_name="power_forecast_curtailment_events",
        schema="performance",
        return_cols=["id"],
        if_exists=if_exists,
    )

    ids = ids_df["id"].to_list()
    logger.debug(f"Inserted {len(ids)} curtailment event(s): {ids}")
    return ids[0] if single_insert else ids

update(event_id=None, description=None, curtailment_percentage=None, start_time=None, end_time=None, user_name=None, data_df=None)

Updates one or more power forecast curtailment events.

You can pass individual values to update a single event, or a DataFrame for batch update.

Parameters:

  • event_id

    (int | None, default: None ) –

    ID of the event to update. Required for single update. By default None.

  • description

    (str | None, default: None ) –

    New description. By default None.

  • curtailment_percentage

    (float | None, default: None ) –

    New curtailment percentage (must be in range (0, 90]). By default None.

  • start_time

    (datetime | None, default: None ) –

    New planned start timestamp. By default None.

  • end_time

    (datetime | None, default: None ) –

    New planned end timestamp. By default None.

  • user_name

    (str | None, default: None ) –

    Name of the user performing the update. By default None.

  • data_df

    (DataFrame | None, default: None ) –

    DataFrame for batch update. Required column: id. Optional: description, curtailment_percentage, start_time, end_time, user_name. When provided, individual parameters are ignored. By default None.

Source code in echo_postgres/kpi_availability_forecast_curtailment_events.py
Python
@validate_call
def update(
    self,
    event_id: int | None = None,
    description: str | None = None,
    curtailment_percentage: float | None = None,
    start_time: datetime | None = None,
    end_time: datetime | None = None,
    user_name: str | None = None,
    data_df: pl.DataFrame | None = None,
) -> None:
    """Updates one or more power forecast curtailment events.

    You can pass individual values to update a single event, or a DataFrame for batch update.

    Parameters
    ----------
    event_id : int | None, optional
        ID of the event to update. Required for single update. By default None.
    description : str | None, optional
        New description. By default None.
    curtailment_percentage : float | None, optional
        New curtailment percentage (must be in range (0, 90]). By default None.
    start_time : datetime | None, optional
        New planned start timestamp. By default None.
    end_time : datetime | None, optional
        New planned end timestamp. By default None.
    user_name : str | None, optional
        Name of the user performing the update. By default None.
    data_df : pl.DataFrame | None, optional
        DataFrame for batch update. Required column: id. Optional: description,
        curtailment_percentage, start_time, end_time, user_name. When provided, individual
        parameters are ignored. By default None.
    """
    df_schema = {
        "id": pl.Int64,
        "description": pl.Utf8,
        "curtailment_percentage": pl.Float64,
        "start_time": pl.Datetime("ms"),
        "end_time": pl.Datetime("ms"),
        "user_name": pl.Utf8,
    }

    if data_df is None:
        single_update = True
        data_df = pl.DataFrame(
            {
                "id": [event_id],
                "description": [description],
                "curtailment_percentage": [curtailment_percentage],
                "start_time": [start_time],
                "end_time": [end_time],
                "user_name": [user_name],
            },
            schema=df_schema,
        )
    else:
        single_update = False
        data_df = data_df.cast({c: t for c, t in df_schema.items() if c in data_df.columns})

    if "id" not in data_df.columns or data_df["id"].is_null().any():
        raise ValueError("'id' column is required and cannot contain nulls.")

    # validate curtailment_percentage range if provided
    if "curtailment_percentage" in data_df.columns:
        invalid_pct = data_df.filter(
            pl.col("curtailment_percentage").is_not_null()
            & ((pl.col("curtailment_percentage") <= 0) | (pl.col("curtailment_percentage") > 90)),
        )
        if len(invalid_pct) > 0:
            raise ValueError("'curtailment_percentage' must be in range (0, 90].")

    # verify IDs exist
    existing = self._perfdb.conn.read_to_polars(
        sql.SQL("SELECT id FROM performance.power_forecast_curtailment_events WHERE id = ANY({ids})").format(
            ids=sql.Literal(data_df["id"].to_list()),
        ),
        schema_overrides={"id": pl.Int64},
    )
    if missing := set(data_df["id"].to_list()) - set(existing["id"].to_list()):
        raise ValueError(f"Curtailment event IDs not found: {missing}")

    # resolve user_name → user_id if provided
    if "user_name" in data_df.columns:
        user_names = data_df["user_name"].drop_nulls().unique().to_list()
        if user_names:
            user_ids = self._perfdb.users.instances.get_ids(names=user_names)
            if missing := set(user_names) - set(user_ids.keys()):
                raise ValueError(f"Users not found: {missing}")
            data_df = data_df.with_columns(
                pl.col("user_name").replace_strict(user_ids, return_dtype=pl.Int64, default=None).alias("user_id"),
            )
        data_df = data_df.drop("user_name")

    self._perfdb.conn.polars_to_sql(
        df=data_df,
        table_name="power_forecast_curtailment_events",
        schema="performance",
        conflict_cols=["id"],
        if_exists="update_only",
        ignore_null_cols=single_update,
    )
    logger.debug(f"Updated {len(data_df)} curtailment event(s).")