-
Notifications
You must be signed in to change notification settings - Fork 0
/
run_resumeEval.py
237 lines (203 loc) · 12.2 KB
/
run_resumeEval.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
import utils.tract as tr
import utils.eval as ev
import utils.vis as vis
import pandas as pd
import json
import os
import matplotlib.pyplot as plt
if __name__ == '__main__':
mode = 'singleExp_afterAnalysis'
if mode == 'singleExp_resumeAnalysis':
# 1.1 norm the prediction
# prototypeArea = getBuildingArea_prototype('C:/Users/xiyu/Downloads/output_data/output_data/EP_output/result_ann_WRF_2018', verbose = 0)
# used for generate the prototype areas, do not have to do it again
experimentLabel = 'energyElec_biLSTM_2023-08-13-02-59-39'
dir_trueTractData = './data/hourly_heat_energy/annual_2016_tract.csv'
target_tractLevel = 'energy.elec'
with open('./saved/estimates_tracts/' + experimentLabel + '/pairListTest.json', 'r') as f:
pairListTest = json.load(f)
dir_buildingMeta = './data/building_metadata/building_metadata.csv'
predPrototypeLevel = tr.reloadPrototypeLevelPred(
'./saved/estimates_tracts/' + experimentLabel + '/buildingLevel')
with open('./data/building_metadata/prototypeArea.json', 'r') as f:
prototypeArea = json.load(f)
prediction_tract_norm = tr.normalize_perM2(predPrototypeLevel, pairListTest, prototypeArea)
# 2.0 get metadata of tracts and remove tracts with the weather not in the test pairs
tractBuildingMeta = tr.getBuildingArea_tracts(dir_buildingMeta, pairListTest)
tractNameRemove = tr.getTracts_remove(dir_buildingMeta, tractBuildingMeta)
tractBuildingMeta = tractBuildingMeta[~tractBuildingMeta['id.tract'].isin(tractNameRemove)]
# 2.1 scaling up
estimate_tract = tr.predict_tracts(prediction_tract_norm, tractBuildingMeta)
# 3 get true data and remove tracts with weather that is not in the test pairs
true_tract = tr.filterTractData(tr.loadTractData(dir_trueTractData, target_tractLevel),
estimate_tract)
true_tract = true_tract[~true_tract.geoid.isin(tractNameRemove)]
# 4 eval
tractsDf = tr.combineEstimateTrue(true_tract, estimate_tract, target_tractLevel)
tractsDf.to_csv('./saved/estimates_tracts/' + experimentLabel + '/' + 'tractsDF.csv')
tr.getTractLevelMetrics(tractsDf,
'./saved/estimates_tracts/' + experimentLabel,
0.1,
)
tr.plotPrototypeLevelMetrics(predPrototypeLevel,
'./saved/estimates_tracts/' + experimentLabel,
ev.cv_mean_absolute_error_wAbs,
'CVMAE_wAbs')
if mode == 'singleExp_afterAnalysis':
######## eval for each experiment
# output the metrics for each prototype
experimentLabel = 'energyElec_biLSTM_10PerData_2023-08-15-16-09-06'
predPrototypeLevel = tr.reloadPrototypeLevelPred(
'./saved/estimates_tracts/' + experimentLabel + '/buildingLevel')
# # get metrics ar prototype
# tr.metricPrototype(tr.metricPrototypeWeather(predPrototypeLevel,ev.cv_mean_absolute_error_wAbs),
# 'CVMAE_wAbs').to_csv('./saved/estimates_tracts/'+ experimentLabel + '/prototypeLevel' + 'CVMAE_wAbs' + '.csv',
# index = False)
# vis.plotPrototypeLevelMetrics_plotly(predPrototypeLevel,
# './saved/estimates_tracts/' + experimentLabel,
# ev.cv_mean_absolute_error_wAbs,
# 'CVMAE_wAbs')
# draw line chart using ground truth and estimation for a prototype
# prototypeSelected = 'HeavyManufacturing-90_1-2004-ASHRAE 169-2013-3B____36'
# prototypeDfSelected = predPrototypeLevel[prototypeSelected]
# prototypeDfSelected['DateTime'] = prototypeDfSelected['DateTime'].apply(lambda x: x.replace('2001', '2018'))
# prototypeDfSelected = prototypeDfSelected.iloc[700: 1440]
# vis.plotPrototypeLevelLines_plotly(prototypeDfSelected,
# './paper/figs/' + 'line_' + prototypeSelected + '.html',
# prototypeSelected.split('_')[0] + ' ' + prototypeSelected.split('_')[-1]
# )
# # draw line chart using ground truth and estimation for a microclimate zone
# # get tract number
# experimentLabel = 'energyElec_biLSTM_10PerData_2023-08-15-16-09-06'
# selectedClimate = [65, 35, 109, 50]
# vis.getTractIndexinWeatherZone(experimentLabel, selectedClimate, './data/building_metadata/building_metadata.csv')
# # line charts
# predTractLevel = pd.read_csv('./saved/estimates_tracts/' + experimentLabel + '/tractsDF.csv')
# vis.plotTractLevelPredictionLine(predTractLevel, 6037600303, './paper/figs/line_tract6037600303.html', [3972,4331])
# # draw census tract level cvmae with std shadow
# experimentLabel = 'energyElec_biLSTM_10PerData_2023-08-15-16-09-06'
# predTractLevel = pd.read_csv('./saved/estimates_tracts/' + experimentLabel + '/tractsDF.csv')
# vis.plotTractLevelCVMAELine_wShadow(predTractLevel,
# # [4332, 5000], # July
# # [6540, 7283], # Oct
# # [8004, 8735], # Dec
# [2148, 2867], # April
# # [0, 8735], # all year
# './paper/figs/line_allTractCVMAEWithShadow_Apr.html')
# # draw census tract level true value with std shadow
# experimentLabel = 'energyElec_biLSTM_10PerData_2023-08-15-16-09-06'
# predTractLevel = pd.read_csv('./saved/estimates_tracts/' + experimentLabel + '/tractsDF.csv')
# vis.plotTractLevelTargetEstimate_wShadow(predTractLevel,
# 'true',
# # [4332, 5000], # July
# # [6540, 7283], # Oct
# # [8004, 8735], # Dec
# # [2148, 2867], # April
# [0, 8735], # all year
# './paper/figs/line_allTractTargetWithShadow.html')
# print peak CVMAE and plot
experimentLabel = 'energyElec_biLSTM_10PerData_2023-08-15-16-09-06'
predTractLevel = pd.read_csv('./saved/estimates_tracts/' + experimentLabel + '/tractsDF.csv')
# vis.plotPeaks_plotly(predTractLevel[predTractLevel.geoid == 6037101122], 'true', './paper/figs/peakValley_6037101122.html')
vis.plotDistPeakCVMAE(predTractLevel, './paper/figs/dist_peakCVMAE.html')
# # output the metrics at tract level
# predTractLevel = pd.read_csv('./saved/estimates_tracts/' + experimentLabel + '/tractsDF.csv')
# tr.getTractLevelMetrics(predTractLevel, './saved/estimates_tracts/' + experimentLabel)
# vis.plotTractLevelPredictionLine(predTractLevel,
# predTractLevel.geoid.sample(1).iloc[0],
# './saved/estimates_tracts/' + experimentLabel)
pass
if mode == 'crossExp':
######### eval across experiments
# prototype level metrics of several targets
# experimentLabels = [
# 'energyElec_biLSTM_10PerData_2023-08-15-16-09-06',
# 'emissionSurf_biLSTM_10PerData_V100_2023-08-17-16-49-19',
# 'emissionExhaust_biLSTM_10PerData_V100_2023-08-17-23-47-42',
# 'emissionRej_biLSTM_10PerData_V100_2023-08-18-02-34-05',
# ]
experimentLabels = [
'energyElec_biLSTM_5PerData_V100_2023-08-17-23-52-29',
'energyElec_biLSTM_10PerData_V100_2023-08-17-20-42-16',
'energyElec_biLSTM_15PerData_V100_2023-08-17-22-11-03',
'energyElec_biLSTM_25PerData_2023-08-15-03-20-09',
]
prototypeLevelMetrics = []
for experiment in experimentLabels:
prototypeLevelMetrics.append(
tr.reloadPrototypeLevelPred('./saved/estimates_tracts/' + experiment + '/buildingLevel'))
vis.plotPrototypeLevelMetrics_plotly(prototypeLevelMetrics,
# [
# 'nMAE_ELEC',
# 'nMAE_E-SURF',
# 'nMAE_E-EXH',
# 'nMAE_E-REJ',
# ],
[
'nMAE_5%',
'nMAE_10%',
'nMAE_15%',
'nMAE_25%',
],
# [
# 'rgb(33,113,181)',
# 'rgb(33,113,181)',
# 'rgb(33,113,181)',
# 'rgb(33,113,181)',
# ],
[
'#BCC6DD',
'#98A3CA',
'#8092C4',
'#455D99',
],
ev.cv_mean_absolute_error_wAbs,
pd.read_csv('./saved/reduced_building_meta/reduced_building_meta.csv'),
'./paper/I3CE',
)
if mode == 'crossExp_prototype_wise_metric_table':
experimentLabels = [
'energyElec_naive_10PerData_2023-08-17-17-02-42',
'energyElec_linear_10PerData_2023-08-17-16-19-02',
'energyElec_mlp_10PerData_2023-08-17-17-13-42',
'energyElec_LSTM_10PerData_2023-08-18-20-04-42',
'energyElec_biLSTM_10PerData_2023-08-15-16-09-06',
]
model_labels = [
'naive', 'linear', 'mlp', 'lstm', 'bilstm'
]
df_list = []
for experiment, mo in zip(experimentLabels, model_labels):
# output the metrics for each prototype
predPrototypeLevel = tr.reloadPrototypeLevelPred(
'./saved/estimates_tracts/' + experiment + '/buildingLevel')
# get metrics at prototype
df = tr.metricPrototype(tr.metricPrototypeWeather(predPrototypeLevel, ev.cv_mean_absolute_error_wAbs),
'CVMAE_wAbs')
df['prototype'] = df['prototype'].str.replace('DOE Ref Pre', '90_1', regex=False).str.strip()
df = df.sort_values('prototype')
df = df.rename(columns={'CVMAE_wAbs': f'{mo}_CVMAE_wAbs'})
df_list.append(df)
df_all = pd.concat(df_list, axis=1)
dup_col = df_all.T.duplicated(keep='first')
df_all = df_all.loc[:, ~dup_col]
if mode == 'others':
######## other test
experimentLabel = 'energyElec_biLSTM_2023-07-20-20-04-01'
predPrototypeLevel = tr.reloadPrototypeLevelPred(
'./saved/estimates_tracts/' + experimentLabel + '/buildingLevel')
df = predPrototypeLevel['SingleFamily-2004____91']
experimentLabel2 = 'energyElec_biLSTM_2023-08-14-03-11-39'
predPrototypeLevel2 = tr.reloadPrototypeLevelPred(
'./saved/estimates_tracts/' + experimentLabel2 + '/buildingLevel')
df2 = predPrototypeLevel2['SingleFamily-2004____91']
df['estimate'].iloc[2100:2300].plot()
df['true'].iloc[2100:2300].plot()
plt.legend()
plt.show()
df2['estimate'].iloc[2100:2300].plot()
df2['true'].iloc[2100:2300].plot()
plt.legend()
plt.show()
print()
print()