def test_save_and_initialise_from_results(example_detector, tmpdir):
det = example_detector
det.run()
evt = EvaluationTool()
evt.build_EvaluationTool_via_run_detector(det)
# Save to HD
evt.store_results_to_HD(tmpdir.join("test_save_evt.txt"))
assert os.path.isfile(tmpdir.join("test_save_evt.txt"))
# Load into a new evt
evt_load = EvaluationTool()
evt_load.build_EvaluationTool_via_results(tmpdir.join("test_save_evt.txt"))
# Check that key quantities have been restored
assert evt_load.type == 4 # Detector has already been run
assert evt_load.results[1] == evt.results[1] # execution time
for rlld_l, rlld in zip(evt_load.results[8],
evt.results[8]): # run length log distribution
assert rlld_l == rlld
"""Fill in and obtain normal and demeaned versions"""
temperatures_spatial[selection, location] = (
station_means[location] + month_means[location, month] +
year_effect_controlling_for_months)
temperatures_spatial_demeaned[selection2, location] = (
temperatures_spatial[selection2, location] -
station_means[location] - month_means[location, month])
"""Normalize"""
temperatures_spatial_demeaned = (temperatures_spatial_demeaned - np.mean(
temperatures_spatial_demeaned, axis=0)) / np.sqrt(
np.var(temperatures_spatial_demeaned, axis=0))
"""STEP 3: Read in the results"""
"""Read in results"""
EvT = EvaluationTool()
EvT.build_EvaluationTool_via_results(results_file)
"""STEP 4: Get your plots"""
segmentation = EvT.results[EvT.names.index("MAP CPs")][-2]
model_labels = EvT.results[EvT.names.index("model labels")]
num_models = len(np.union1d(model_labels, model_labels))
relevant_models = np.union1d([seg[1] for seg in segmentation],
[seg[1] for seg in segmentation])
mods = [8, 11, 13, 17, 18]
all_models = np.linspace(0,
len(model_labels) - 1,
len(model_labels),
dtype=int)
""""STEP 5: Get annotations"""
#http://file.scirp.org/pdf/ACS_2013062615184222.pdf
#https://en.wikipedia.org/wiki/History_of_climate_change_science
#https://en.wikipedia.org/wiki/Climate_change_in_Europe
reader = csv.reader(csvfile)
for row in reader:
raw_data += row
raw_data_float = []
for entry in raw_data:
raw_data_float.append(float(entry))
raw_data = raw_data_float
T = int(len(raw_data))
S1, S2 = 1,1 #S1, S2 give you spatial dimensions
data = np.array(raw_data).reshape(T,1,1)
"""STEP 2: Read in the data to create your EvT objects"""
EvTKL = EvaluationTool()
EvTKL.build_EvaluationTool_via_results(result_path + "//" + results_file_KL)
EvTDPD = EvaluationTool()
EvTDPD.build_EvaluationTool_via_results(result_path + "//" + results_file_DPD)
if plot1:
"""STEP 3: Set up the figure properties and plot"""
height_ratio =[8,10] #10,10
custom_colors = ["blue", "purple"]
fig, ((ax1, ax2),(ax3,ax4)) = plt.subplots(2, 2,
figsize=(12,5),
sharex = 'col',
sharey = 'row',# True,
gridspec_kw = {'height_ratios':height_ratio})
ax_arrays = np.array([np.array([ax1, ax2]), np.array([ax3, ax4])])
int(len(model_universe)))
detector = Detector(
data, model_universe, model_prior, cp_model,
S1, S2, T, exo_data=None, num_exo_vars=None,
threshold=200,
store_rl=True, store_mrl=True)#,
detector.run()
"""build evaluation tool"""
EvT = EvaluationTool()
EvT.build_EvaluationTool_via_run_detector(detector)
else:
"""Directly read results"""
result_path = baseline_working_directory
EvT = EvaluationTool()
EvT.build_EvaluationTool_via_results(result_path + "//" + "results_demo.txt")
"""Plot in panels: Raw data (offsetting the mean),
CPs + RLD
1-step-ahead-prediction + variance
model posterior"""
custom_colors_models = ['green', 'purple', 'orange', 'blue', 'darkgray']
custom_colors_series = ['black']*4
custom_linestyles =['solid']*5
offsets = np.linspace(1,S1*S2, S1*S2).reshape(S1, S2)
data_original = data.copy()
data = data + offsets*2.5
data = data.reshape(T, S1*S2)
""""Get some quantities needed for rest"""
with open(baseline_working_directory + "//" + date_file) as csvfile:
reader = csv.reader(csvfile)
for row in reader:
#DEBUG: Unclear if this is needed
if count > 0:
myl += row
count += 1
if count % 2500 == 0:
print(count)
dates = []
for e in myl:
dates.append(int(e))
result_path = baseline_working_directory
EvT = EvaluationTool()
EvT.build_EvaluationTool_via_results(result_path + "//" + results_file)
"""Using the dates, select your range and indices: select
03/07/1975 -- 31/12/2008, i.e. find indices that correspond"""
start_test = dates.index(19740703)
start_algo = dates.index(19750703)
stop = dates.index(20081231)
"""time period for which we want RLD"""
start_comparison = dates.index(20070801)#dates.index(19980102)
stop_comparison = stop#stop
all_dates = []
for d in dates[start_comparison:stop-2]:
s = str(d)
all_dates.append(datetime.date(year=int(s[0:4]), month=int(s[4:6]), day=int(s[6:8])))
for _2h in range(0, 12 * 7):
selection_2h = [False] * _2h + [
True
] + [False] * (12 * 7 - 1 - _2h)
selection = (selection_2h * int(T / (7 * 12)) +
selection_2h[:(T - int(T / (7 * 12)) * 7 * 12)])
mean_2h[_2h, station] = np.mean(data[selection, station])
data[selection, station] = (data[selection, station] -
mean_2h[_2h, station])
if normalize:
data = (data - np.mean(data, axis=0)) / np.sqrt(np.var(data, axis=0))
intercept_priors = np.mean(data, axis=0)
"""""STEP 2: READ RESULTS""" ""
EvTKL, EvTDPD = EvaluationTool(), EvaluationTool()
EvTKL.build_EvaluationTool_via_results(results_file_KL)
EvTDPD.build_EvaluationTool_via_results(results_file_DPD)
"""Get dates"""
def perdelta(start, end, delta, date_list):
curr = start
while curr < end:
#yield curr
date_list.append(curr)
curr += delta
all_dates = []
#start_year, start_month, start_day, start_hour = 2002, 8, 17, 0
#start_datetime = datetime.datetime(year = 2002, month = 8, day = 17, hour = 0)