Mirrors/EOS
1
0
Fork 0
mirror of https://github.com/Akkudoktor-EOS/EOS.git synced 2025-12-22 11:26:20 +00:00
Code Issues Actions 1 Packages Projects Releases Wiki Activity

fix: load data for automatic optimization (#731)

Browse Source 
Automatic optimization used to take the adjusted load data even if there were no
measurements leading to 0 load values.

Split LoadAkkudoktor into LoadAkkudoktor and LoadAkkudoktorAdjusted. This allows
to select load data either purely from the load data database or load data additionally
adjusted by load measurements. Some value names have been adapted to denote
also the unit of a value.

For better load bug squashing the optimization solution data availability was
improved. For better data visbility prediction data can now be distinguished from
solution data in the generic optimization solution.

Some predictions that may be of interest to understand the solution were added.

Documentation was updated to resemble the addition load prediction provider and
the value name changes.

Signed-off-by: Bobby Noelte <b0661n0e17e@gmail.com>
This commit is contained in:
Bobby Noelte
2025-11-01 00:49:11 +01:00
committed by GitHub
parent e3c5b758dd
commit b01bb1c61c
26 changed files with 515 additions and 227 deletions
Download Patch File Download Diff File Expand all files Collapse all files

188
src/akkudoktoreos/server/dash/plan.py
View File

@@ -7,6 +7,7 @@ from bokeh.plotting import figure
from loguru import logger
from monsterui.franken import (
Card,
CardTitle,
Details,
Div,
DivLAligned,
@@ -33,10 +34,33 @@ from akkudoktoreos.utils.datetimeutil import compare_datetimes, to_datetime
# bar width for 1 hour bars (time given in millseconds)
BAR_WIDTH_1HOUR = 1000 * 60 * 60
# Tailwind compatible color palette
color_palette = {
"red-500": "#EF4444", # red-500
"orange-500": "#F97316", # orange-500
"amber-500": "#F59E0B", # amber-500
"yellow-500": "#EAB308", # yellow-500
"lime-500": "#84CC16", # lime-500
"green-500": "#22C55E", # green-500
"emerald-500": "#10B981", # emerald-500
"teal-500": "#14B8A6", # teal-500
"cyan-500": "#06B6D4", # cyan-500
"sky-500": "#0EA5E9", # sky-500
"blue-500": "#3B82F6", # blue-500
"indigo-500": "#6366F1", # indigo-500
"violet-500": "#8B5CF6", # violet-500
"purple-500": "#A855F7", # purple-500
"pink-500": "#EC4899", # pink-500
"rose-500": "#F43F5E", # rose-500
}
colors = list(color_palette.keys())
# Current state of solution displayed
solution_visible: dict[str, bool] = {
"pv_prediction_energy_wh": True,
"elec_price_prediction_amt_kwh": True,
"pv_energy_wh": True,
"elec_price_amt_kwh": True,
"feed_in_tariff_amt_kwh": True,
}
solution_color: dict[str, str] = {}
@@ -75,6 +99,7 @@ def SolutionCard(solution: OptimizationSolution, config: SettingsEOS, data: Opti
Args:
data (Optional[dict]): Incoming data containing action and category for processing.
"""
global colors, color_palette
category = "solution"
dark = False
if data and data.get("category", None) == category:
@@ -86,11 +111,34 @@ def SolutionCard(solution: OptimizationSolution, config: SettingsEOS, data: Opti
if data and data.get("dark", None) == "true":
dark = True
df = solution.data.to_dataframe()
df = solution.solution.to_dataframe()
if df.empty or len(df.columns) <= 1:
raise ValueError(f"DataFrame is empty or missing plottable columns: {list(df.columns)}")
raise ValueError(
f"Solution DataFrame is empty or missing plottable columns: {list(df.columns)}"
)
if "date_time" not in df.columns:
raise ValueError(f"DataFrame is missing column 'date_time': {list(df.columns)}")
raise ValueError(f"Solution DataFrame is missing column 'date_time': {list(df.columns)}")
solution_columns = list(df.columns)
instruction_columns = [
instruction
for instruction in solution_columns
if instruction.endswith("op_mode") or instruction.endswith("op_factor")
]
solution_columns = [x for x in solution_columns if x not in instruction_columns]
prediction_df = solution.prediction.to_dataframe()
if prediction_df.empty or len(prediction_df.columns) <= 1:
raise ValueError(
f"Prediction DataFrame is empty or missing plottable columns: {list(prediction_df.columns)}"
)
if "date_time" not in prediction_df.columns:
raise ValueError(
f"Prediction DataFrame is missing column 'date_time': {list(prediction_df.columns)}"
)
prediction_columns = list(prediction_df.columns)
prediction_columns_to_join = prediction_df.columns.difference(df.columns)
df = df.join(prediction_df[prediction_columns_to_join], how="inner")
# Remove time offset from UTC to get naive local time and make bokey plot in local time
dst_offsets = df.index.map(lambda x: x.dst().total_seconds() / 3600)
@@ -192,7 +240,6 @@ def SolutionCard(solution: OptimizationSolution, config: SettingsEOS, data: Opti
# Create line renderers for each column
renderers = {}
colors = ["black", "blue", "cyan", "green", "orange", "pink", "purple"]
for i, col in enumerate(sorted(df.columns)):
# Exclude some columns that are currently not used or are covered by others
@@ -218,24 +265,24 @@ def SolutionCard(solution: OptimizationSolution, config: SettingsEOS, data: Opti
solution_visible[col] = visible
if col in solution_color:
color = solution_color[col]
elif col == "pv_prediction_energy_wh":
color = "yellow"
elif col == "pv_energy_wh":
color = "yellow-500"
solution_color[col] = color
elif col == "elec_price_prediction_amt_kwh":
color = "red"
elif col == "elec_price_amt_kwh":
color = "red-500"
solution_color[col] = color
else:
color = colors[i % len(colors)]
solution_color[col] = color
if visible:
if col == "pv_prediction_energy_wh":
if col == "pv_energy_wh":
r = plot.vbar(
x="date_time",
top=col,
source=source,
width=BAR_WIDTH_1HOUR * 0.8,
legend_label=col,
color=color,
color=color_palette[color],
level="underlay",
)
elif col.endswith("energy_wh"):
@@ -245,7 +292,7 @@ def SolutionCard(solution: OptimizationSolution, config: SettingsEOS, data: Opti
mode="before",
source=source,
legend_label=col,
color=color,
color=color_palette[color],
)
elif col.endswith("factor"):
r = plot.step(
@@ -254,7 +301,7 @@ def SolutionCard(solution: OptimizationSolution, config: SettingsEOS, data: Opti
mode="before",
source=source,
legend_label=col,
color=color,
color=color_palette[color],
y_range_name="factor",
)
elif col.endswith("mode"):
@@ -264,7 +311,7 @@ def SolutionCard(solution: OptimizationSolution, config: SettingsEOS, data: Opti
mode="before",
source=source,
legend_label=col,
color=color,
color=color_palette[color],
y_range_name="factor",
)
elif col.endswith("amt_kwh"):
@@ -274,7 +321,7 @@ def SolutionCard(solution: OptimizationSolution, config: SettingsEOS, data: Opti
mode="before",
source=source,
legend_label=col,
color=color,
color=color_palette[color],
y_range_name="amt_kwh",
)
elif col.endswith("amt"):
@@ -284,7 +331,7 @@ def SolutionCard(solution: OptimizationSolution, config: SettingsEOS, data: Opti
mode="before",
source=source,
legend_label=col,
color=color,
color=color_palette[color],
y_range_name="amt",
)
else:
@@ -298,34 +345,93 @@ def SolutionCard(solution: OptimizationSolution, config: SettingsEOS, data: Opti
# --- CheckboxGroup to toggle datasets ---
Checkbox = Grid(
*[
LabelCheckboxX(
label=renderer,
id=f"{renderer}-visible",
name=f"{renderer}-visible",
value="true",
checked=solution_visible[renderer],
hx_post="/eosdash/plan",
hx_target="#page-content",
hx_swap="innerHTML",
hx_vals='js:{ "category": "solution", "action": "visible", "renderer": '
+ '"'
+ f"{renderer}"
+ '", '
+ '"dark": window.matchMedia("(prefers-color-scheme: dark)").matches '
+ "}",
lbl_cls=f"text-{solution_color[renderer]}-500",
)
for renderer in list(renderers.keys())
],
cols=2,
Card(
Grid(
*[
LabelCheckboxX(
label=renderer,
id=f"{renderer}-visible",
name=f"{renderer}-visible",
value="true",
checked=solution_visible[renderer],
hx_post="/eosdash/plan",
hx_target="#page-content",
hx_swap="innerHTML",
hx_vals='js:{ "category": "solution", "action": "visible", "renderer": '
+ '"'
+ f"{renderer}"
+ '", '
+ '"dark": window.matchMedia("(prefers-color-scheme: dark)").matches '
+ "}",
lbl_cls=f"text-{solution_color[renderer]}",
)
for renderer in list(renderers.keys())
if renderer in prediction_columns
],
cols=2,
),
header=CardTitle("Prediction"),
),
Card(
Grid(
*[
LabelCheckboxX(
label=renderer,
id=f"{renderer}-visible",
name=f"{renderer}-visible",
value="true",
checked=solution_visible[renderer],
hx_post="/eosdash/plan",
hx_target="#page-content",
hx_swap="innerHTML",
hx_vals='js:{ "category": "solution", "action": "visible", "renderer": '
+ '"'
+ f"{renderer}"
+ '", '
+ '"dark": window.matchMedia("(prefers-color-scheme: dark)").matches '
+ "}",
lbl_cls=f"text-{solution_color[renderer]}",
)
for renderer in list(renderers.keys())
if renderer in solution_columns
],
cols=2,
),
header=CardTitle("Solution"),
),
Card(
Grid(
*[
LabelCheckboxX(
label=renderer,
id=f"{renderer}-visible",
name=f"{renderer}-visible",
value="true",
checked=solution_visible[renderer],
hx_post="/eosdash/plan",
hx_target="#page-content",
hx_swap="innerHTML",
hx_vals='js:{ "category": "solution", "action": "visible", "renderer": '
+ '"'
+ f"{renderer}"
+ '", '
+ '"dark": window.matchMedia("(prefers-color-scheme: dark)").matches '
+ "}",
lbl_cls=f"text-{solution_color[renderer]}",
)
for renderer in list(renderers.keys())
if renderer in instruction_columns
],
cols=2,
),
header=CardTitle("Instruction"),
),
cols=1,
)
return Grid(
Bokeh(plot),
Card(
Checkbox,
),
Checkbox,
cls="w-full space-y-3 space-x-3",
)

24
src/akkudoktoreos/server/dash/prediction.py
View File

@@ -153,9 +153,9 @@ def WeatherIrradianceForecast(
def LoadForecast(predictions: pd.DataFrame, config: dict, date_time_tz: str, dark: bool) -> FT:
source = ColumnDataSource(predictions)
provider = config["load"]["provider"]
if provider == "LoadAkkudoktor":
if provider == "LoadAkkudoktorAdjusted":
year_energy = config["load"]["provider_settings"]["LoadAkkudoktor"][
"loadakkudoktor_year_energy"
"loadakkudoktor_year_energy_kwh"
]
provider = f"{provider}, {year_energy} kWh"
@@ -168,8 +168,8 @@ def LoadForecast(predictions: pd.DataFrame, config: dict, date_time_tz: str, dar
height=400,
)
# Add secondary y-axis for stddev
stddev_min = predictions["load_std"].min()
stddev_max = predictions["load_std"].max()
stddev_min = predictions["loadakkudoktor_std_power_w"].min()
stddev_max = predictions["loadakkudoktor_std_power_w"].max()
plot.extra_y_ranges["stddev"] = Range1d(start=stddev_min - 5, end=stddev_max + 5)
y2_axis = LinearAxis(y_range_name="stddev", axis_label="Load Standard Deviation [W]")
y2_axis.axis_label_text_color = "green"
@@ -177,21 +177,21 @@ def LoadForecast(predictions: pd.DataFrame, config: dict, date_time_tz: str, dar
plot.line(
"date_time",
"load_mean",
"loadforecast_power_w",
source=source,
legend_label="Load mean value",
legend_label="Load forcast value (adjusted by measurement)",
color="red",
)
plot.line(
"date_time",
"load_mean_adjusted",
"loadakkudoktor_mean_power_w",
source=source,
legend_label="Load adjusted by measurement",
legend_label="Load mean value",
color="blue",
)
plot.line(
"date_time",
"load_std",
"loadakkudoktor_std_power_w",
source=source,
legend_label="Load standard deviation",
color="green",
@@ -233,9 +233,9 @@ def Prediction(eos_host: str, eos_port: Union[str, int], data: Optional[dict] =
"weather_ghi",
"weather_dni",
"weather_dhi",
"load_mean",
"load_std",
"load_mean_adjusted",
"loadforecast_power_w",
"loadakkudoktor_std_power_w",
"loadakkudoktor_mean_power_w",
],
}
result = requests.get(f"{server}/v1/prediction/dataframe", params=params, timeout=10)

14
src/akkudoktoreos/server/eos.py
View File

@@ -1243,7 +1243,7 @@ async def fastapi_gesamtlast(request: GesamtlastRequest) -> list[float]:
filled with the first available prediction value.
Note:
Use '/v1/prediction/list?key=load_mean_adjusted' instead.
Use '/v1/prediction/list?key=loadforecast_power_w' instead.
Load energy meter readings to be added to EOS measurement by:
'/v1/measurement/value' or
'/v1/measurement/series' or
@@ -1255,10 +1255,10 @@ async def fastapi_gesamtlast(request: GesamtlastRequest) -> list[float]:
"hours": request.hours,
},
"load": {
"provider": "LoadAkkudoktor",
"provider": "LoadAkkudoktorAdjusted",
"provider_settings": {
"LoadAkkudoktor": {
"loadakkudoktor_year_energy": request.year_energy,
"loadakkudoktor_year_energy_kwh": request.year_energy,
},
},
},
@@ -1317,7 +1317,7 @@ async def fastapi_gesamtlast(request: GesamtlastRequest) -> list[float]:
end_datetime = start_datetime.add(days=2)
try:
prediction_list = prediction_eos.key_to_array(
key="load_mean_adjusted",
key="loadforecast_power_w",
start_datetime=start_datetime,
end_datetime=end_datetime,
).tolist()
@@ -1347,14 +1347,14 @@ async def fastapi_gesamtlast_simple(year_energy: float) -> list[float]:
Set LoadAkkudoktor as provider, then update data with
'/v1/prediction/update'
and then request data with
'/v1/prediction/list?key=load_mean' instead.
'/v1/prediction/list?key=loadforecast_power_w' instead.
"""
settings = SettingsEOS(
load=LoadCommonSettings(
provider="LoadAkkudoktor",
provider_settings=LoadCommonProviderSettings(
LoadAkkudoktor=LoadAkkudoktorCommonSettings(
loadakkudoktor_year_energy=year_energy / 1000, # Convert to kWh
loadakkudoktor_year_energy_kwh=year_energy / 1000, # Convert to kWh
),
),
)
@@ -1378,7 +1378,7 @@ async def fastapi_gesamtlast_simple(year_energy: float) -> list[float]:
end_datetime = start_datetime.add(days=2)
try:
prediction_list = prediction_eos.key_to_array(
key="load_mean",
key="loadforecast_power_w",
start_datetime=start_datetime,
end_datetime=end_datetime,
).tolist()