# Copyright (c) 2021. Bram Vonk, Enexis import pandas as pd [docs]def format_limits(df_meta, df_data=None): """ Format the transformer limits for plotting. Parameters ---------- df_meta: pd.DataFrame DataFrame with the trasnformer limits. df_data: pd.DataFrame Optionally, the historic preprocess to determine the time range for the limits. Returns ------- A DataFrame with the limits to plot. """ limit = df_meta["vermogen_nominaal"].squeeze() if df_data is None: return pd.DataFrame({"extreme": ["min", "max"], "value": [-limit, limit]}) else: df_plot = df_data.sort_values(["date"]).iloc[[1, -1]][["date"]] df_plot["lower"] = -limit df_plot["upper"] = limit return df_plot.melt(id_vars=["date"], value_vars=["lower", "upper"], var_name=["limit"]) [docs]def format_history(df_data): """ Format the historic for plotting. Parameters ---------- df_data: pd.DataFrame The historic preprocess to be plotted. Returns ------- A DataFrame with the historic preprocess to plot. """ df_plot = df_data.copy() value_vars = ["min", "max"] df_plot = df_plot.melt( var_name="history", value_name="power", id_vars=df_plot.columns.difference(value_vars), value_vars=value_vars, ) return df_plot [docs]def format_forecast(df_forecast): """ Format the forecasted load. Parameters ---------- df: pd.DataFrame Returns ------- pass through input """ return df_forecast [docs]def dummy_forecast(df): """ Create a dummy forecast with median and Q10-Q90 based on historic preprocess to test plotting function. Parameters ---------- df : pd.DataFrame Historic preprocess to use as a basis Returns ------- pd.DataFrame Forecast preprocess that can be used to plot with plot_forecast() """ extremes = ["max", "min"] df = df.melt( id_vars=df.columns.difference(extremes), value_vars=extremes, var_name="extreme", ) df["forecast"] = "Q10-Q90" df["upper"] = df["value"] + 50 df["lower"] = df["value"] - 50 df_median = df.copy() df_median["forecast"] = "median" df_median["upper"] = df_median["value"] df_median["lower"] = df_median["value"] df_forecast = pd.concat([df, df_median]) return df_forecast