def test_resampling_argument_value():
arg = (
(ts_as_data_frame, "Q", "mean"),
(ts_as_series, "D", "median"),
(ts_as_data_frame.tz_localize("UTC").tz_convert("US/Pacific"), "W", "sum"),
)
for a in arg:
with pytest.raises(ValueError):
resample_time_series(a[0], a[1], agg=a[2])
def test_daily_resampling_incomplete_and_time_shift_mean():
start = pd.Timestamp(2016, 1, 1, 18)
end = pd.Timestamp(2016, 1, 31, 18)
arg = (
ts_as_data_frame[start:end].tz_localize("UTC").tz_convert("ETC/GMT+8"),
ts_as_series[start:end].tz_localize("UTC").tz_convert("ETC/GMT-8"),
)
for i, a in enumerate(arg):
ts_resampled = resample_time_series(a, "D", agg="mean")
assert len(a) != len(ts_resampled)
if i == 0:
# first day (2016/01/01) and last day (2016/01/31) are incomplete.
# 31 days are available in January 2016 when resampling using mean.
assert len(ts_resampled) == 31
assert str(ts_resampled.index[0]) == str(
pd.Timestamp("2016-01-01", tz="ETC/GMT+8")
)
assert str(ts_resampled.index[-1]) == str(
pd.Timestamp("2016-01-31", tz="ETC/GMT+8")
)
else:
# first day (2016/01/02) and last day (2016/02/01) are incomplete.
# 30 days are available in January 2016 and 1 day available in February 2016
assert len(ts_resampled) == 31
assert str(ts_resampled.index[0]) == str(
pd.Timestamp("2016-01-02", tz="ETC/GMT-8")
)
assert str(ts_resampled.index[-1]) == str(
pd.Timestamp("2016-02-01", tz="ETC/GMT-8")
)
def test_weekly_resampling_incomplete_and_time_shift_mean():
start = pd.Timestamp(2016, 1, 1, 18)
end = pd.Timestamp(2016, 1, 31, 18)
arg = (
ts_as_data_frame[start:end].tz_localize("UTC").tz_convert("ETC/GMT+8"),
ts_as_series[start:end].tz_localize("UTC").tz_convert("ETC/GMT-8"),
)
for i, a in enumerate(arg):
ts_resampled = resample_time_series(a, "W", agg="mean")
assert len(a) != len(ts_resampled)
assert len(ts_resampled) == 6
if i == 0:
# first week (2015/12/27) and last week (2016/01/31) are incomplete.
# 1 week is available in 2015 and 5 weeks are available in 2016 when
# resampling using mean.
assert str(ts_resampled.index[0]) == str(
pd.Timestamp("2015-12-27", tz="ETC/GMT+8")
)
assert str(ts_resampled.index[-1]) == str(
pd.Timestamp("2016-01-31", tz="ETC/GMT+8")
)
else:
# first week (2015/12/27) and last week (2016/01/31) are incomplete.
# 1 week is available in 2015 and 5 weeks are available in 2016 when
# resampling using mean.
assert str(ts_resampled.index[0]) == str(
pd.Timestamp("2015-12-27", tz="ETC/GMT-8")
)
assert str(ts_resampled.index[-1]) == str(
pd.Timestamp("2016-01-31", tz="ETC/GMT-8")
)
def test_daily_resampling_time_shift_mean():
arg = (
ts_as_data_frame.tz_localize("UTC").tz_convert("ETC/GMT+8"),
ts_as_series.tz_localize("UTC").tz_convert("ETC/GMT-8"),
)
for i, a in enumerate(arg):
ts_resampled = resample_time_series(a, "D", agg="mean")
assert len(a) != len(ts_resampled)
assert len(ts_resampled) == 367
if i == 0:
# first day (2015/12/31) and last day (2016/12/31) are incomplete.
# 1 day is available in 2015 and 366 days are available in 2016
assert str(ts_resampled.index[0]) == str(
pd.Timestamp("2015-12-31", tz="ETC/GMT+8")
)
assert str(ts_resampled.index[-1]) == str(
pd.Timestamp("2016-12-31", tz="ETC/GMT+8")
)
assert ts_resampled.iloc[0, 0] == ts_resampled.iloc[-1, 0] == 1
assert ts_resampled.iloc[0, 1] == ts_resampled.iloc[-1, 1] == 2
else:
# first day (2016/01/01) and last day (2017/01/01) are incomplete.
# 366 days are available in 2016 and 1 day is available in 2017
assert str(ts_resampled.index[0]) == str(
pd.Timestamp("2016-01-01", tz="ETC/GMT-8")
)
assert str(ts_resampled.index[-1]) == str(
pd.Timestamp("2017-01-01", tz="ETC/GMT-8")
)
assert ts_resampled.iloc[0] == ts_resampled.iloc[-1] == 1
def test_monthly_resampling_time_shift_mean():
arg = (
ts_as_data_frame.tz_localize("UTC").tz_convert("ETC/GMT+8"),
ts_as_series.tz_localize("UTC").tz_convert("ETC/GMT-8"),
)
for i, a in enumerate(arg):
ts_resampled = resample_time_series(a, "M", agg="mean")
assert len(a) != len(ts_resampled)
if i == 0:
# first month (2015/12) and last month (2016/12) are incomplete.
# 1 month is available in 2015 and 12 months are available in 2016 after
# resampling using mean
assert len(ts_resampled) == 13
assert str(ts_resampled.index[0]) == str(
pd.Timestamp("2015-12-01", tz="ETC/GMT+8")
)
assert str(ts_resampled.index[-1]) == str(
pd.Timestamp("2016-12-01", tz="ETC/GMT+8")
)
else:
# first month (2016/01) and last month (2017/01) are incomplete.
# 12 months are available in 2016 and 1 month available in 2017 after
# resampling using mean
assert len(ts_resampled) == 13
assert str(ts_resampled.index[0]) == str(
pd.Timestamp("2016-01-01", tz="ETC/GMT-8")
)
assert str(ts_resampled.index[-1]) == str(
pd.Timestamp("2017-01-01", tz="ETC/GMT-8")
)
def test_weekly_resampling_incomplete_and_time_shift_sum():
start = pd.Timestamp(2016, 1, 1, 18)
end = pd.Timestamp(2016, 1, 31, 18)
arg = (
ts_as_data_frame[start:end].tz_localize("UTC").tz_convert("ETC/GMT+8"),
ts_as_series[start:end].tz_localize("UTC").tz_convert("ETC/GMT-8"),
)
for i, a in enumerate(arg):
ts_resampled = resample_time_series(a, "W")
assert len(a) != len(ts_resampled)
if i == 0:
# first week (2015/12/27) and last week (2016/01/31) are incomplete.
# 4 full weeks are available in 2016 after resampling using sum.
assert len(ts_resampled) == 4
assert str(ts_resampled.index[0]) == str(
pd.Timestamp("2016-01-03", tz="ETC/GMT+8")
)
assert str(ts_resampled.index[-1]) == str(
pd.Timestamp("2016-01-24", tz="ETC/GMT+8")
)
else:
# first week (2015/12/27) and last week (2016/01/31) are incomplete.
# 4 full weeks are available in 2016 after resampling using sum.
assert len(ts_resampled) == 4
assert str(ts_resampled.index[0]) == str(
pd.Timestamp("2016-01-03", tz="ETC/GMT-8")
)
assert str(ts_resampled.index[-1]) == str(
pd.Timestamp("2016-01-24", tz="ETC/GMT-8")
)
def test_daily_resampling_time_shift_sum():
arg = (
ts_as_data_frame.tz_localize("UTC").tz_convert("ETC/GMT+8"),
ts_as_series.tz_localize("UTC").tz_convert("ETC/GMT-8"),
)
for i, a in enumerate(arg):
ts_resampled = resample_time_series(a, "D")
assert len(a) != len(ts_resampled)
if i == 0:
# first day (2015/12/31) and last day (2016/12/31) are incomplete.
# 365 full days are available in 2016 after resampling using sum.
assert len(ts_resampled) == 365
assert str(ts_resampled.index[0]) == str(
pd.Timestamp("2016-01-01", tz="ETC/GMT+8")
)
assert str(ts_resampled.index[-1]) == str(
pd.Timestamp("2016-12-30", tz="ETC/GMT+8")
)
else:
# first day (2016/01/01) and last day (2017/01/01) are incomplete.
# 365 full days are available in 2016 after resampling using sum.
assert len(ts_resampled) == 365
assert str(ts_resampled.index[0]) == str(
pd.Timestamp("2016-01-02", tz="ETC/GMT-8")
)
assert str(ts_resampled.index[-1]) == str(
pd.Timestamp("2016-12-31", tz="ETC/GMT-8")
)
def test_weekly_resampling_time_shift_mean():
arg = (
ts_as_data_frame.tz_localize("UTC").tz_convert("ETC/GMT+8"),
ts_as_series.tz_localize("UTC").tz_convert("ETC/GMT-8"),
)
for i, a in enumerate(arg):
ts_resampled = resample_time_series(a, "W", agg="mean")
assert len(a) != len(ts_resampled)
if i == 0:
# first week (2015/12/27) and last week (2016/12/25) are incomplete.
# 1 week is available in 2015 and 52 weeks are available in 2016
assert len(ts_resampled) == 53
assert str(ts_resampled.index[0]) == str(
pd.Timestamp("2015-12-27", tz="ETC/GMT+8")
)
assert str(ts_resampled.index[-1]) == str(
pd.Timestamp("2016-12-25", tz="ETC/GMT+8")
)
else:
# first week (2015/12/27) and last week (2017/01/01) are incomplete.
# 1 week is available in 2015, 52 weeks are available in 2016 and 1 week
# is available in 2017
assert len(ts_resampled) == 54
assert str(ts_resampled.index[0]) == str(
pd.Timestamp("2015-12-27", tz="ETC/GMT-8")
)
assert str(ts_resampled.index[-1]) == str(
pd.Timestamp("2017-01-01", tz="ETC/GMT-8")
)
def test_monthly_resampling_incomplete_and_time_shift_mean():
start = pd.Timestamp(2016, 1, 15, 18)
end = pd.Timestamp(2016, 6, 30, 18)
arg = (
ts_as_data_frame[start:end].tz_localize("UTC").tz_convert("ETC/GMT+8"),
ts_as_series[start:end].tz_localize("UTC").tz_convert("ETC/GMT-8"),
)
for i, a in enumerate(arg):
ts_resampled = resample_time_series(a, "M", agg="mean")
assert len(a) != len(ts_resampled)
if i == 0:
# first month (2016/01) and last month (2016/06) are incomplete.
# 6 months are available in 2016 after resampling using mean.
assert len(ts_resampled) == 6
assert str(ts_resampled.index[0]) == str(
pd.Timestamp("2016-01-01", tz="ETC/GMT+8")
)
assert str(ts_resampled.index[-1]) == str(
pd.Timestamp("2016-06-01", tz="ETC/GMT+8")
)
else:
# first month (2016/01) and last month (2016/07) are incomplete.
# 7 months are available in 2016 after resampling using mean.
assert len(ts_resampled) == 7
assert str(ts_resampled.index[0]) == str(
pd.Timestamp("2016-01-01", tz="ETC/GMT-8")
)
assert str(ts_resampled.index[-1]) == str(
pd.Timestamp("2016-07-01", tz="ETC/GMT-8")
)
def test_weekly_resampling_time_shift_sum():
arg = (
ts_as_data_frame.tz_localize("UTC").tz_convert("ETC/GMT+8"),
ts_as_series.tz_localize("UTC").tz_convert("ETC/GMT-8"),
)
for i, a in enumerate(arg):
ts_resampled = resample_time_series(a, "W")
assert len(a) != len(ts_resampled)
if i == 0:
# first week (2015/12/28) and last week (2016/12/25) are incomplete.
# 51 full weeks are available in 2016 after resampling using sum.
assert len(ts_resampled) == 51
assert str(ts_resampled.index[0]) == str(
pd.Timestamp("2016-01-03", tz="ETC/GMT+8")
)
assert str(ts_resampled.index[-1]) == str(
pd.Timestamp("2016-12-18", tz="ETC/GMT+8")
)
else:
# first week (2015/12/28) is incomplete.
# 52 full weeks are available in 2016 after resampling using sum.
assert len(ts_resampled) == 52
assert str(ts_resampled.index[0]) == str(
pd.Timestamp("2016-01-03", tz="ETC/GMT-8")
)
assert str(ts_resampled.index[-1]) == str(
pd.Timestamp("2016-12-25", tz="ETC/GMT-8")
)
def test_monthly_resampling_sum():
arg = (ts_as_data_frame, ts_as_series)
for i, a in enumerate(arg):
ts_resampled = resample_time_series(a, "M")
assert len(a) != len(ts_resampled)
# 12 full monts are available in 2016 when resampling using mean.
assert len(ts_resampled) == 12
assert str(ts_resampled.index[1]) == str(pd.Timestamp(2016, 2, 1))
if i == 0:
assert a.sum().equals(ts_resampled.sum())
else:
assert a.sum() == ts_resampled.sum()
def test_weekly_resampling_sum():
arg = (ts_as_data_frame, ts_as_series)
for i, a in enumerate(arg):
ts_resampled = resample_time_series(a, "W")
assert len(a) != len(ts_resampled)
# first week (2015/12/28) is incomplete.
# 52 full weeks are available in 2016 after resampling using sum.
assert len(ts_resampled) == 52
assert ts_resampled.index.dayofweek[0] == 6 # week starts on Sunday
if i == 0:
assert not a.sum().equals(ts_resampled.sum())
else:
assert a.sum() != ts_resampled.sum()
def test_daily_resampling_mean():
arg = (ts_as_data_frame, ts_as_series)
for i, a in enumerate(arg):
ts_resampled = resample_time_series(a, "D", agg="mean")
assert len(a) != len(ts_resampled)
# 366 days are available in 2016
assert len(ts_resampled) == 366
assert str(ts_resampled.index[1]) == str(pd.Timestamp(2016, 1, 2))
if i == 0:
assert ts_resampled.iloc[0, 0] == ts_resampled.iloc[-1, 0] == 1
assert ts_resampled.iloc[0, 1] == ts_resampled.iloc[-1, 1] == 2
else:
assert ts_resampled.iloc[0] == ts_resampled.iloc[-1] == 1
def test_monthly_resampling_mean():
arg = (ts_as_data_frame, ts_as_series)
for i, a in enumerate(arg):
ts_resampled = resample_time_series(a, "M", agg="mean")
assert len(a) != len(ts_resampled)
# 12 full months are available in 2016 when resampling using mean.
assert len(ts_resampled) == 12
assert str(ts_resampled.index[2]) == str(pd.Timestamp(2016, 3, 1))
if i == 0:
assert ts_resampled.iloc[0, 0] == ts_resampled.iloc[-1, 0] == 1
assert ts_resampled.iloc[0, 1] == ts_resampled.iloc[-1, 1] == 2
else:
assert ts_resampled.iloc[0] == ts_resampled.iloc[-1] == 1
def test_daily_resampling_sum():
arg = (ts_as_data_frame, ts_as_series)
for i, a in enumerate(arg):
ts_resampled = resample_time_series(a, "D")
assert len(a) != len(ts_resampled)
# 366 days are available in 2016
assert len(ts_resampled) == 366
assert str(ts_resampled.index[0]) == str(pd.Timestamp(2016, 1, 1))
assert str(ts_resampled.index[-1]) == str(pd.Timestamp(2016, 12, 31))
if i == 0:
assert a.sum().equals(ts_resampled.sum())
else:
assert a.sum() == ts_resampled.sum()
def test_monthly_resampling_incomplete_mean():
start = pd.Timestamp(2016, 1, 15)
end = pd.Timestamp(2016, 12, 15)
arg = (ts_as_data_frame, ts_as_series)
for i, a in enumerate(arg):
ts_resampled = resample_time_series(a[start:end], "M", agg="mean")
# first month (01/2016) and last month (2016/12) are incomplete.
# 12 months are available in 2016 when resampling using mean.
assert len(ts_resampled) == 12
assert str(ts_resampled.index[0]) == str(pd.Timestamp(2016, 1, 1))
if i == 0:
assert ts_resampled.iloc[0, 0] == ts_resampled.iloc[-1, 0] == 1
assert ts_resampled.iloc[0, 1] == ts_resampled.iloc[-1, 1] == 2
else:
assert ts_resampled.iloc[0] == ts_resampled.iloc[-1] == 1
def test_monthly_resampling_incomplete_sum():
start = pd.Timestamp(2016, 1, 15)
end = pd.Timestamp(2016, 12, 15)
arg = (ts_as_data_frame, ts_as_series)
for i, a in enumerate(arg):
ts_resampled = resample_time_series(a[start:end], "M")
# first month (01/2016) and last month (2016/12) are incomplete.
# 10 full months are available in 2016 when resampling using sum.
assert len(a[start:end]) != len(ts_resampled)
assert len(ts_resampled) == 10 # inncomplete months are clipped
assert str(ts_resampled.index[0]) == str(pd.Timestamp(2016, 2, 1))
if i == 0:
assert not a.sum().equals(ts_resampled.sum())
else:
assert a.sum() != ts_resampled.sum()
def test_weekly_resampling_mean():
arg = (ts_as_data_frame, ts_as_series)
for i, a in enumerate(arg):
ts_resampled = resample_time_series(a, "W", agg="mean")
assert len(a) != len(ts_resampled)
# first week (2015/12/27) is incomplete.
# 1 week is available in 2015 and 52 weeks are available in 2016
assert len(ts_resampled) == 53
assert ts_resampled.index.dayofweek[0] == 6 # week starts on Sunday
assert ts_resampled.index[0] == pd.Timestamp(2015, 12, 27)
assert ts_resampled.index[-1] == pd.Timestamp(2016, 12, 25)
if i == 0:
assert ts_resampled.iloc[0, 0] == ts_resampled.iloc[-1, 0] == 1
assert ts_resampled.iloc[0, 1] == ts_resampled.iloc[-1, 1] == 2
else:
assert ts_resampled.iloc[0] == ts_resampled.iloc[-1] == 1
def test_daily_resampling_incomplete_mean():
start = pd.Timestamp(2016, 1, 1, 12)
end = pd.Timestamp(2016, 1, 31, 12)
arg = (ts_as_data_frame[start:end], ts_as_series[start:end])
for i, a in enumerate(arg):
ts_resampled = resample_time_series(a, "D", agg="mean")
# first day (2016/01/01) and last day (2016/01/31) are incomplete.
# 31 days are available in January 2016 after resampling using mean.
assert len(ts_resampled) == 31
assert str(ts_resampled.index[0]) == str(pd.Timestamp(2016, 1, 1))
assert str(ts_resampled.index[-1]) == str(pd.Timestamp(2016, 1, 31))
if i == 0:
assert ts_resampled.iloc[0, 0] == ts_resampled.iloc[-1, 0] == 1
assert ts_resampled.iloc[0, 1] == ts_resampled.iloc[-1, 1] == 2
else:
assert ts_resampled.iloc[0] == ts_resampled.iloc[-1] == 1
def test_weekly_resampling_incomplete_mean():
start = pd.Timestamp(2016, 1, 1, 12)
end = pd.Timestamp(2016, 1, 31, 12)
arg = (ts_as_data_frame[start:end], ts_as_series[start:end])
for i, a in enumerate(arg):
ts_resampled = resample_time_series(a, "W", agg="mean")
# first week (2015/12/27) and last week (2016/01/31) are incomplete.
# 1 week is available in 2015 and 5 weeks are available in 2016 after
# resampling using mean.
assert len(ts_resampled) == 6
assert str(ts_resampled.index[0]) == str(pd.Timestamp(2015, 12, 27))
assert str(ts_resampled.index[-1]) == str(pd.Timestamp(2016, 1, 31))
if i == 0:
assert ts_resampled.iloc[0, 0] == ts_resampled.iloc[-1, 0] == 1
assert ts_resampled.iloc[0, 1] == ts_resampled.iloc[-1, 1] == 2
else:
assert ts_resampled.iloc[0] == ts_resampled.iloc[-1] == 1
def test_daily_resampling_incomplete_sum():
start = pd.Timestamp(2016, 1, 1, 12)
end = pd.Timestamp(2016, 1, 31, 12)
arg = (ts_as_data_frame[start:end], ts_as_series[start:end])
for i, a in enumerate(arg):
ts_resampled = resample_time_series(a[start:end], "D")
assert len(a) != len(ts_resampled)
# first day (2016/01/01) and last day (2016/01/31) are incomplete.
# 29 days are available in Jamuary 2016 after resampling using sum.
assert len(ts_resampled) == 29
assert str(ts_resampled.index[0]) == str(pd.Timestamp(2016, 1, 2))
if i == 0:
assert ts_resampled.sum().equals(
pd.Series([float(29 * 24), float(29 * 2 * 24)], index=["A", "B"])
)
else:
assert ts_resampled.sum() == 29 * 24
def test_weekly_resampling_incomplete_sum():
start = pd.Timestamp(2016, 1, 1, 12)
end = pd.Timestamp(2016, 1, 31, 12)
arg = (ts_as_data_frame[start:end], ts_as_series[start:end])
for i, a in enumerate(arg):
ts_resampled = resample_time_series(a[start:end], "W")
assert len(a) != len(ts_resampled)
# first week (2015/12/28) and last week (2016/01/31) are incomplete.
# 4 full weeks are available in 2016 after resampling using sum.
assert len(ts_resampled) == 4
assert str(ts_resampled.index[0]) == str(pd.Timestamp(2016, 1, 3))
assert str(ts_resampled.index[-1]) == str(pd.Timestamp(2016, 1, 24))
if i == 0:
assert ts_resampled.sum().equals(
pd.Series([float(4 * 7 * 24), float(4 * 7 * 2 * 24)], index=["A", "B"])
)
else:
assert ts_resampled.sum() == 4 * 7 * 24
def plot_generation_time_series_stack(
scenario,
area,
resources,
area_type=None,
time_range=None,
time_zone="utc",
time_freq="H",
show_demand=True,
show_net_demand=True,
normalize=False,
t2c=None,
t2l=None,
t2hc=None,
title=None,
label_fontsize=20,
title_fontsize=22,
tick_fontsize=15,
legend_fontsize=18,
save=False,
filename=None,
filepath=None,
):
"""Generate time series generation stack plot in a certain area of a scenario.
:param powersimdata.scenario.scenario.Scenario scenario: scenario instance
:param str area: one of *loadzone*, *state*, *state abbreviation*,
*interconnect*, *'all'*
:param str/list resources: one or a list of resources. *'solar_curtailment'*,
*'wind_curtailment'*, *'wind_offshore_curtailment'* are valid entries together
with all available generator types in the area. The order of the resources
determines the stack order in the figure.
:param str area_type: one of *'loadzone'*, *'state'*, *'state_abbr'*,
*'interconnect'*
:param tuple time_range: [start_timestamp, end_timestamp] where each time stamp
is pandas.Timestamp/numpy.datetime64/datetime.datetime. If None, the entire
time range is used for the given scenario.
:param str time_zone: new time zone.
:param str time_freq: frequency. Either *'D'* (day), *'W'* (week), *'M'* (month).
:param bool show_demand: show demand line in the plot or not, default is True.
:param bool show_net_demand: show net demand line in the plot or not, default is
True.
:param bool normalize: normalize the generation based on capacity or not,
default is False.
:param dict t2c: user specified color of resource type to overwrite type2color
default dict. key: resource type, value: color code.
:param dict t2l: user specified label of resource type to overwrite type2label
default dict. key: resource type, value: label.
:param dict t2hc: user specified color of curtailable resource hatches to overwrite
type2hatchcolor default dict. key: resource type, valid keys are
*'wind_curtailment'*, *'solar_curtailment'*, *'wind_offshore_curtailment'*,
value: color code.
:param str title: user specified title of the figure, default is set to be area.
:param float label_fontsize: user specified label fontsize, default is 20.
:param float title_fontsize: user specified title fontsize, default is 22.
:param float tick_fontsize: user specified ticks of axes fontsize, default is 15.
:param float legend_fontsize: user specified legend fontsize, default is 18.
:param bool save: save the generated figure or not, default is False.
:param str filename: if save is True, user specified filename, use area if None.
:param str filepath: if save is True, user specified filepath, use current
directory if None.
"""
_check_scenario_is_in_analyze_state(scenario)
mi = ModelImmutables(scenario.info["grid_model"])
type2color = mi.plants["type2color"]
type2label = mi.plants["type2label"]
type2hatchcolor = mi.plants["type2hatchcolor"]
if t2c:
type2color.update(t2c)
if t2l:
type2label.update(t2l)
if t2hc:
type2hatchcolor.update(t2hc)
pg_stack = get_generation_time_series_by_resources(scenario,
area,
resources,
area_type=area_type)
capacity = get_capacity_by_resources(scenario,
area,
resources,
area_type=area_type)
demand = get_demand_time_series(scenario, area, area_type=area_type)
net_demand = get_net_demand_time_series(scenario,
area,
area_type=area_type)
capacity_ts = pd.Series(capacity.sum(), index=pg_stack.index)
curtailable_resources = {
"solar_curtailment",
"wind_curtailment",
"wind_offshore_curtailment",
}
if curtailable_resources & set(resources):
curtailment = get_curtailment_time_series(scenario,
area,
area_type=area_type)
for r in curtailable_resources:
if r in resources and r in curtailment.columns:
pg_stack[r] = curtailment[r]
if time_zone != "utc":
pg_stack = change_time_zone(pg_stack, time_zone)
demand = change_time_zone(demand, time_zone)
net_demand = change_time_zone(net_demand, time_zone)
capacity_ts = change_time_zone(capacity_ts, time_zone)
if not time_range:
time_range = (
pd.Timestamp(scenario.info["start_date"]),
pd.Timestamp(scenario.info["end_date"]),
)
pg_stack = slice_time_series(pg_stack, time_range[0], time_range[1])
demand = slice_time_series(demand, time_range[0], time_range[1])
net_demand = slice_time_series(net_demand, time_range[0], time_range[1])
capacity_ts = slice_time_series(capacity_ts, time_range[0], time_range[1])
if time_freq != "H":
pg_stack = resample_time_series(pg_stack, time_freq)
demand = resample_time_series(demand, time_freq)
net_demand = resample_time_series(net_demand, time_freq)
capacity_ts = resample_time_series(capacity_ts, time_freq)
if "storage" in resources:
pg_storage = get_storage_time_series(scenario,
area,
area_type=area_type)
capacity_storage = get_storage_capacity(scenario,
area,
area_type=area_type)
capacity_storage_ts = pd.Series(capacity_storage,
index=pg_storage.index)
if time_zone != "utc":
pg_storage = change_time_zone(pg_storage, time_zone)
capacity_storage_ts = change_time_zone(capacity_storage_ts,
time_zone)
if time_range != ("2016-01-01 00:00:00", "2016-12-31 23:00:00"):
pg_storage = slice_time_series(pg_storage, time_range[0],
time_range[1])
capacity_storage_ts = slice_time_series(capacity_storage_ts,
time_range[0],
time_range[1])
if time_freq != "H":
pg_storage = resample_time_series(pg_storage, time_freq)
capacity_storage_ts = resample_time_series(capacity_storage_ts,
time_freq)
pg_stack["storage"] = pg_storage.clip(lower=0)
capacity_ts += capacity_storage_ts
fig, (ax, ax_storage) = plt.subplots(
2,
1,
figsize=(20, 15),
sharex="row",
gridspec_kw={
"height_ratios": [3, 1],
"hspace": 0.02
},
)
plt.subplots_adjust(wspace=0)
if normalize:
pg_storage = pg_storage.divide(capacity_storage_ts, axis="index")
ax_storage.set_ylabel("Normalized Storage",
fontsize=label_fontsize)
else:
ax_storage.set_ylabel("Energy Storage (MW)",
fontsize=label_fontsize)
ax_storage = pg_storage.plot(color=type2color["storage"],
lw=4,
ax=ax_storage)
ax_storage.fill_between(
pg_storage.index.values,
0,
pg_storage.values,
color=type2color["storage"],
alpha=0.5,
)
# Erase year in xticklabels
xt_with_year = list(ax_storage.__dict__["date_axis_info"][0])
xt_with_year[-1] = b"%b"
ax_storage.__dict__["date_axis_info"][0] = tuple(xt_with_year)
ax_storage.tick_params(axis="both",
which="both",
labelsize=tick_fontsize)
ax_storage.set_xlabel("")
for a in fig.get_axes():
a.label_outer()
else:
fig = plt.figure(figsize=(20, 10))
ax = fig.gca()
if normalize:
pg_stack = pg_stack.divide(capacity_ts, axis="index")
demand = demand.divide(capacity_ts, axis="index")
net_demand = net_demand.divide(capacity_ts, axis="index")
ax.set_ylabel("Normalized Generation", fontsize=label_fontsize)
else:
pg_stack = pg_stack.divide(1e6, axis="index")
demand = demand.divide(1e6, axis="index")
net_demand = net_demand.divide(1e6, axis="index")
ax.set_ylabel("Daily Energy TWh", fontsize=label_fontsize)
available_resources = [r for r in resources if r in pg_stack.columns]
pg_stack[available_resources].clip(0, None).plot.area(color=type2color,
linewidth=0,
alpha=0.7,
ax=ax,
sharex="row")
if show_demand:
demand.plot(color="red", lw=4, ax=ax)
if show_net_demand:
net_demand.plot(color="red", ls="--", lw=2, ax=ax)
if not title:
title = area
ax.set_title("%s" % title, fontsize=title_fontsize)
ax.grid(color="black", axis="y")
if "storage" not in resources:
# Erase year in xticklabels
xt_with_year = list(ax.__dict__["date_axis_info"][0])
xt_with_year[-1] = b"%b"
ax.__dict__["date_axis_info"][0] = tuple(xt_with_year)
ax.set_xlabel("")
ax.tick_params(which="both", labelsize=tick_fontsize)
ax.set_ylim([
min(0, 1.1 * net_demand.min()),
max(ax.get_ylim()[1], 1.1 * demand.max()),
])
handles, labels = ax.get_legend_handles_labels()
if show_demand:
labels[0] = "Demand"
if show_net_demand:
labels[1] = "Net Demand"
label_offset = show_demand + show_net_demand
labels = [type2label[l] if l in type2label else l for l in labels]
# Add hatches
for r in curtailable_resources:
if r in available_resources:
ind = available_resources.index(r)
ax.fill_between(
pg_stack[available_resources].index.values,
pg_stack[available_resources].iloc[:, :ind + 1].sum(axis=1),
pg_stack[available_resources].iloc[:, :ind].sum(axis=1),
color="none",
hatch="//",
edgecolor=type2hatchcolor[r],
linewidth=0.0,
)
handles[ind + label_offset] = mpatches.Patch(
facecolor=type2color[r],
hatch="//",
edgecolor=type2hatchcolor[r],
linewidth=0.0,
)
ax.legend(
handles[::-1],
labels[::-1],
frameon=2,
prop={"size": legend_fontsize},
loc="upper left",
bbox_to_anchor=(1, 1),
)
if save:
if not filename:
filename = area
if not filepath:
filepath = os.path.join(os.getcwd(), filename)
plt.savefig(f"{filepath}.pdf", bbox_inches="tight", pad_inches=0)
def plot_curtailment_time_series(
scenario,
area,
resources,
area_type=None,
time_range=None,
time_zone="utc",
time_freq="H",
show_demand=True,
percentage=True,
t2c=None,
t2l=None,
title=None,
label_fontsize=20,
title_fontsize=22,
tick_fontsize=15,
legend_fontsize=18,
save=False,
filename=None,
filepath=None,
):
"""Generate time series curtailment plot of each specified resource in a certain
area of a scenario.
:param powersimdata.scenario.scenario.Scenario scenario: scenario instance
:param str area: one of *loadzone*, *state*, *state abbreviation*,
*interconnect*, *'all'*
:param str/list resources: one or a list of resources.
:param str area_type: one of *'loadzone'*, *'state'*, *'state_abbr'*,
*'interconnect'*
:param tuple time_range: [start_timestamp, end_timestamp] where each time stamp
is pandas.Timestamp/numpy.datetime64/datetime.datetime. If None, the entire
time range is used for the given scenario.
:param str time_zone: new time zone.
:param str time_freq: frequency. Either *'D'* (day), *'W'* (week), *'M'* (month).
:param bool show_demand: show demand line in the plot or not, default is True.
:param bool percentage: plot the curtailment in terms of percentage or not,
default is True.
:param dict t2c: user specified color of resource type to overwrite type2color
default dict. key: resource type, value: color code.
:param dict t2l: user specified label of resource type to overwrite type2label
default dict. key: resource type, value: label.
:param str title: user specified title of the figure.
:param float label_fontsize: user specified label fontsize, default is 20.
:param float title_fontsize: user specified title fontsize, default is 22.
:param float tick_fontsize: user specified ticks of axes fontsize, default is 15.
:param float legend_fontsize: user specified legend fontsize, default is 18.
:param bool save: save the generated figure or not, default is False.
:param str filename: if save is True, user specified filename, use area if None.
:param str filepath: if save is True, user specified filepath, use current
directory if None.
"""
_check_scenario_is_in_analyze_state(scenario)
resources = _check_resources_and_format(
resources, grid_model=scenario.info["grid_model"])
mi = ModelImmutables(scenario.info["grid_model"])
type2color = mi.plants["type2color"]
type2label = mi.plants["type2label"]
if t2c:
type2color.update(t2c)
if t2l:
type2label.update(t2l)
resource_pg = get_generation_time_series_by_resources(scenario,
area,
resources,
area_type=area_type)
demand = get_demand_time_series(scenario, area, area_type=area_type)
curtailment = get_curtailment_time_series(scenario,
area,
area_type=area_type)
resource_pg = change_time_zone(resource_pg, time_zone)
demand = change_time_zone(demand, time_zone)
curtailment = change_time_zone(curtailment, time_zone)
if not time_range:
time_range = (
pd.Timestamp(scenario.info["start_date"], tz="utc"),
pd.Timestamp(scenario.info["end_date"], tz="utc"),
)
resource_pg = slice_time_series(resource_pg, time_range[0], time_range[1])
demand = slice_time_series(demand, time_range[0], time_range[1])
curtailment = slice_time_series(curtailment, time_range[0], time_range[1])
if time_freq != "H":
resource_pg = resample_time_series(resource_pg, time_freq)
demand = resample_time_series(demand, time_freq)
curtailment = resample_time_series(curtailment, time_freq)
for r in resource_pg.columns:
curtailment[r + "_curtailment" +
"_mean"] = curtailment[r + "_curtailment"].mean()
curtailment[
r +
"_available"] = curtailment[r + "_curtailment"] + resource_pg[r]
curtailment[r + "_curtailment" +
"_percentage"] = (curtailment[r + "_curtailment"] /
curtailment[r + "_available"] * 100)
curtailment[r + "_curtailment" + "_percentage" +
"_mean"] = curtailment[r + "_curtailment" +
"_percentage"].mean()
for r in resources:
if r not in resource_pg.columns:
raise ValueError(f"{r} is invalid in {area}!")
fig = plt.figure(figsize=(20, 10))
ax = fig.gca()
title_text = f"{area} {r.capitalize()}" if not title else title
plt.title(title_text, fontsize=title_fontsize)
cr = r + "_curtailment"
if percentage:
key1, key2 = f"{cr}_percentage", f"{cr}_percentage_mean"
else:
key1, key2 = cr, f"{cr}_mean"
curtailment[key1].plot(
ax=ax,
lw=4,
alpha=0.7,
color=type2color[cr],
label=type2label[cr],
)
curtailment[key2].plot(
ax=ax,
ls="--",
lw=4,
alpha=0.7,
color=type2color[cr],
label=type2label[cr] + " Mean",
)
ax_twin = ax.twinx()
curtailment[r + "_available"].plot(
ax=ax_twin,
lw=4,
alpha=0.7,
color=type2color[r],
label=f"{type2label[r]} Energy Available",
)
if show_demand:
demand.plot(ax=ax_twin,
lw=4,
alpha=0.7,
color="red",
label="Demand")
# Erase year in xticklabels
xt_with_year = list(ax_twin.__dict__["date_axis_info"][0])
xt_with_year[-1] = b"%b"
ax_twin.__dict__["date_axis_info"][0] = tuple(xt_with_year)
ax_twin.set_xlabel("")
ax_twin.tick_params(which="both", labelsize=tick_fontsize)
ax_twin.yaxis.get_offset_text().set_fontsize(tick_fontsize)
ax_twin.set_ylabel("MWh", fontsize=label_fontsize)
ax_twin.legend(loc="upper right", prop={"size": legend_fontsize})
ax.tick_params(which="both", labelsize=tick_fontsize)
ax.yaxis.get_offset_text().set_fontsize(tick_fontsize)
ax.grid(color="black", axis="y")
ax.set_xlabel("")
if percentage:
ax.set_ylabel("Curtailment [%]", fontsize=label_fontsize)
else:
ax.set_ylabel("Curtailment", fontsize=label_fontsize)
ax.legend(loc="upper left", prop={"size": legend_fontsize})
if save:
if not filename:
filename = f"{area.lower()}_{r}_curtailment"
if not filepath:
filepath = os.path.join(os.getcwd(), filename)
plt.savefig(f"{filepath}.pdf", bbox_inches="tight", pad_inches=0)
