Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

CODS follow-up #333

Merged
merged 6 commits into from
Sep 13, 2022
Merged

CODS follow-up #333

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
6 commits
Select commit Hold shift + click to select a range
bc55fce
detection_tuner -> cleaning_sensitivity
kandersolar Jul 6, 2022
89ef905
pruning_tuner -> clean_pruning_sensitivity
kandersolar Jul 6, 2022
99bb095
fix address pandas deprecation warning
kandersolar Jul 6, 2022
ae4edbb
remove warnings from cods notebook
kandersolar Jul 6, 2022
42333e3
docstring corrections
kandersolar Jul 6, 2022
7e68242
add PR number to pending.rst entry
kandersolar Aug 1, 2022
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
25 changes: 2 additions & 23 deletions docs/cods_example.ipynb
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -336,29 +336,8 @@
"15 False \n",
"\n",
"Bootstrapping for uncertainty analysis (16 realizations):\n",
"# 16 | Used: 0.4 min | Left: 0.0 min | Progress: [----------------------------->] 100 %"
]
},
{
"name": "stderr",
"output_type": "stream",
"text": [
"/Users/mdecegli/opt/anaconda3/envs/cods_test/lib/python3.10/site-packages/rdtools/soiling.py:1875: FutureWarning: In a future version of pandas all arguments of concat except for the argument 'objs' will be keyword-only\n",
" concat_tot_mod = pd.concat([kdf.total_model for kdf in bt_kdfs], 1)\n",
"/Users/mdecegli/opt/anaconda3/envs/cods_test/lib/python3.10/site-packages/rdtools/soiling.py:1876: FutureWarning: In a future version of pandas all arguments of concat except for the argument 'objs' will be keyword-only\n",
" concat_SR = pd.concat([kdf.soiling_ratio for kdf in bt_kdfs], 1)\n",
"/Users/mdecegli/opt/anaconda3/envs/cods_test/lib/python3.10/site-packages/rdtools/soiling.py:1877: FutureWarning: In a future version of pandas all arguments of concat except for the argument 'objs' will be keyword-only\n",
" concat_r_s = pd.concat([kdf.soiling_rates for kdf in bt_kdfs], 1)\n",
"/Users/mdecegli/opt/anaconda3/envs/cods_test/lib/python3.10/site-packages/rdtools/soiling.py:1878: FutureWarning: In a future version of pandas all arguments of concat except for the argument 'objs' will be keyword-only\n",
" concat_ce = pd.concat([kdf.cleaning_events for kdf in bt_kdfs], 1)\n"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\n",
"Final RMSE: 0.01723\n"
"# 16 | Used: 0.5 min | Left: 0.0 min | Progress: [----------------------------->] 100 %\n",
"Final RMSE: 0.01722\n"
]
}
],
Expand Down
2 changes: 1 addition & 1 deletion docs/sphinx/source/changelog/pending.rst
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -8,7 +8,7 @@ Enhancements
now shows the number of year-on-year slopes in addition to color coding points
(:issue:`298`, :pull:`324`)
* The Combined estimation Of Degradation and Soiling (CODS) algorithm is implemented
in the soiling module and illustrated in an example notebook (:pull:`150`)
in the soiling module and illustrated in an example notebook (:pull:`150`, :pull:`333`)
* Circular block bootstrapping added as a method for calculating uncertainty in
``degradation_year_on_year()`` via the ``Uncertainty_method`` argument (:pull:`150`)

Expand Down
79 changes: 37 additions & 42 deletions rdtools/soiling.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1154,9 +1154,6 @@ class CODSAnalysis():
adf_res : list
The results of an Augmented Dickey-Fuller test (telling whether the
residuals are stationary or not)
_parameters_n_weights : pandas.DataFrame
Contains information about the parameters used in each bootstrap model
fit, and the resultant weight.

Raises
------
Expand All @@ -1183,8 +1180,8 @@ def __init__(self, energy_normalized_daily):

def iterative_signal_decomposition(
self, order=('SR', 'SC', 'Rd'), degradation_method='YoY',
max_iterations=18, detection_tuner=.5, convergence_criterion=5e-3,
pruning_iterations=1, pruning_tuner=.6, soiling_significance=.75,
max_iterations=18, cleaning_sensitivity=.5, convergence_criterion=5e-3,
pruning_iterations=1, clean_pruning_sensitivity=.6, soiling_significance=.75,
process_noise=1e-4, renormalize_SR=None, ffill=True, clip_soiling=True,
verbose=False):
'''
Expand Down Expand Up @@ -1226,15 +1223,15 @@ def iterative_signal_decomposition(
max_iterations : int, default 18
Max number of iterations to perform. Each iteration fits only one of the
components, so three iterations are needed to fit all three components.
detection_tuner : float, default .5
cleaning_sensitivity : float, default .5
Higher value gives lower cleaning event detection sensitivity.
Should be between 0.1 and 2
convergence_criterion : float, default 5e-3
the relative change in the convergence metric required for
convergence
pruning_iterations : int, default 1
Number of iterations when pruning (removing) cleaning events
pruning_tuner : float, default .6
clean_pruning_sensitivity : float, default .6
Sensitivity tuner that decides how easily a cleaning event is pruned
(removed). Larger values means a smaller chance of pruning a given event.
Should be between 0.1 and 2
Expand Down Expand Up @@ -1293,13 +1290,11 @@ def iterative_signal_decomposition(
+------------------------+----------------------------------------------+
| Key | Description |
+========================+==============================================+
| 'degradation' | List of linear degradation rate of system in |
| | %/year, lower and upper bound of 95% |
| | confidence interval (list) |
| 'degradation' | Linear degradation rate of system in %/year |
| | (float) |
+------------------------+----------------------------------------------+
| 'soiling_loss' | List of average soiling losses over the time |
| | series in %, lower and upper bound of 95% |
| | confidence interval (list) |
| 'soiling_loss' | Average soiling losses over the time series |
| | in % (float) |
+------------------------+----------------------------------------------+
| 'residual_shift' | Mean value of residuals. Multiply total |
| | model by this number for complete overlap |
Expand Down Expand Up @@ -1349,7 +1344,7 @@ def iterative_signal_decomposition(

# Find possible cleaning events based on the performance index
ce, rm9 = _rolling_median_ce_detection(pi.index, pi, ffill=ffill,
tuner=detection_tuner)
tuner=cleaning_sensitivity)
pce = _collapse_cleaning_events(ce, rm9.diff().values, 5)

small_soiling_signal, perfect_cleaning = False, True
Expand All @@ -1368,13 +1363,13 @@ def iterative_signal_decomposition(
if ic > 2: # Add possible cleaning events found by considering
# the residuals
pce = soiling_dfs[-1].cleaning_events.copy()
detection_tuner *= 1.2 # Increase value of detection tuner
cleaning_sensitivity *= 1.2 # decrease sensitivity
ce, rm9 = _rolling_median_ce_detection(
pi.index, residuals, ffill=ffill,
tuner=detection_tuner)
tuner=cleaning_sensitivity)
ce = _collapse_cleaning_events(ce, rm9.diff().values, 5)
pce[ce] = True
pruning_tuner /= 1.1 # Decrease value of pruning tuner
clean_pruning_sensitivity /= 1.1 # increase pruning sensitivity

# Decompose input signal
soiling_dummy = (pi /
Expand All @@ -1388,7 +1383,7 @@ def iterative_signal_decomposition(
clip_soiling=clip_soiling,
prescient_cleaning_events=pce,
pruning_iterations=pruning_iterations,
pruning_tuner=pruning_tuner,
clean_pruning_sensitivity=clean_pruning_sensitivity,
perfect_cleaning=perfect_cleaning,
process_noise=process_noise,
renormalize_SR=renormalize_SR)
Expand Down Expand Up @@ -1542,8 +1537,8 @@ def run_bootstrap(self,
process_noise=1e-4,
order_alternatives=(('SR', 'SC', 'Rd'),
('SC', 'SR', 'Rd')),
detection_tuner_alternatives=(.25, .75),
pruning_tuner_alternatives=(1/1.5, 1.5),
cleaning_sensitivity_alternatives=(.25, .75),
clean_pruning_sensitivity_alternatives=(1/1.5, 1.5),
forward_fill_alternatives=(True, False),
verbose=False,
**kwargs):
Expand Down Expand Up @@ -1571,7 +1566,7 @@ def run_bootstrap(self,
Number of bootstrap realizations to be run
minimum N, where N is the possible combinations of model
alternatives defined by possible combination of order_alternatives,
detection_tuner_alternatives, pruning_tuner_alternatives and
cleaning_sensitivity_alternatives, clean_pruning_sensitivity_alternatives and
forward_fill_alternatives.
confidence_level : float, default 68.2
The size of the confidence intervals to return, in percent
Expand All @@ -1583,11 +1578,11 @@ def run_bootstrap(self,
order_alternatives : tuple of tuples, default (('SR', 'SC', 'Rd'), ('SC', 'SR', 'Rd'))
Component estimation orders that will be tested during initial
model fitting.
detection_tuner_alternatives : tuple, default (.25, .75)
cleaning_sensitivity_alternatives : tuple, default (.25, .75)
Detection tuner values that will be tested during initial fitting.
Length must be >= 1. First and last values define limits of values
that will be used during bootstrapping.
pruning_tuner_alternatives : tuple, default (1/1.5, 1.5)
clean_pruning_sensitivity_alternatives : tuple, default (1/1.5, 1.5)
Pruning tuner values that will be tested during initial fitting.
Length must be >= 1. First and last values define limits of values
that will be used during bootstrapping.
Expand Down Expand Up @@ -1675,8 +1670,8 @@ def run_bootstrap(self,

# Generate combinations of model parameter alternatives
parameter_alternatives = [order_alternatives,
detection_tuner_alternatives,
pruning_tuner_alternatives,
cleaning_sensitivity_alternatives,
clean_pruning_sensitivity_alternatives,
forward_fill_alternatives]
index_list = list(itertools.product([0, 1], repeat=len(parameter_alternatives)))
combination_of_parameters = [[parameter_alternatives[j][indexes[j]]
Expand All @@ -1697,8 +1692,8 @@ def run_bootstrap(self,
try:
df_out, result_dict = self.iterative_signal_decomposition(
max_iterations=18, order=order, clip_soiling=True,
detection_tuner=dt, pruning_iterations=1,
pruning_tuner=pt, process_noise=process_noise, ffill=ff,
cleaning_sensitivity=dt, pruning_iterations=1,
clean_pruning_sensitivity=pt, process_noise=process_noise, ffill=ff,
degradation_method=degradation_method, **kwargs)

# Save results
Expand Down Expand Up @@ -1824,8 +1819,8 @@ def run_bootstrap(self,
# Do Signal decomposition for soiling and degradation component
kdf, results_dict = temporary_cods_instance.iterative_signal_decomposition(
max_iterations=4, order=order, clip_soiling=True,
detection_tuner=dt, pruning_iterations=1,
pruning_tuner=pt, process_noise=pn,
cleaning_sensitivity=dt, pruning_iterations=1,
clean_pruning_sensitivity=pt, process_noise=pn,
renormalize_SR=renormalize_SR, ffill=ffill,
degradation_method=degradation_method, **kwargs)

Expand Down Expand Up @@ -1872,10 +1867,10 @@ def run_bootstrap(self,
ci_high_edge = (50 + confidence_level / 2) / 100

# Concatenate boostrap model fits
concat_tot_mod = pd.concat([kdf.total_model for kdf in bt_kdfs], 1)
concat_SR = pd.concat([kdf.soiling_ratio for kdf in bt_kdfs], 1)
concat_r_s = pd.concat([kdf.soiling_rates for kdf in bt_kdfs], 1)
concat_ce = pd.concat([kdf.cleaning_events for kdf in bt_kdfs], 1)
concat_tot_mod = pd.concat([kdf.total_model for kdf in bt_kdfs], axis=1)
concat_SR = pd.concat([kdf.soiling_ratio for kdf in bt_kdfs], axis=1)
concat_r_s = pd.concat([kdf.soiling_rates for kdf in bt_kdfs], axis=1)
concat_ce = pd.concat([kdf.cleaning_events for kdf in bt_kdfs], axis=1)

# Find confidence intervals for SR and soiling rates
df_out['SR_low'] = concat_SR.quantile(ci_low_edge, 1)
Expand Down Expand Up @@ -1935,7 +1930,7 @@ def run_bootstrap(self,

def _Kalman_filter_for_SR(self, zs_series, process_noise=1e-4, zs_std=.05,
rate_std=.005, max_soiling_rates=.0005,
pruning_iterations=1, pruning_tuner=.6,
pruning_iterations=1, clean_pruning_sensitivity=.6,
renormalize_SR=None, perfect_cleaning=False,
prescient_cleaning_events=None,
clip_soiling=True, ffill=True):
Expand All @@ -1960,7 +1955,7 @@ def _Kalman_filter_for_SR(self, zs_series, process_noise=1e-4, zs_std=.05,
Represents the maximum allowed positive soiling rate (when soiling is removed)
pruning_iterations : int, default 1
Number of iterations when pruning (removing) cleaning events
pruning_tuner : float, default 0.6
clean_pruning_sensitivity : float, default 0.6
Sensitivity tuner that decides how easily a cleaning event is pruned
(removed). Larger values means a smaller chance of pruning a given event.
renormalize_SR : float or None, default None
Expand Down Expand Up @@ -2093,7 +2088,7 @@ def _Kalman_filter_for_SR(self, zs_series, process_noise=1e-4, zs_std=.05,
pi_after_cleaning = rm_smooth_pi.loc[cleaning_events]
# Detect outiers/false positives
false_positives = _find_numeric_outliers(pi_after_cleaning,
pruning_tuner, 'lower')
clean_pruning_sensitivity, 'lower')
cleaning_events = \
false_positives[~false_positives].index.tolist()

Expand Down Expand Up @@ -2280,8 +2275,8 @@ def soiling_cods(energy_normalized_daily,
process_noise=1e-4,
order_alternatives=(('SR', 'SC', 'Rd'),
('SC', 'SR', 'Rd')),
detection_tuner_alternatives=(.25, .75),
pruning_tuner_alternatives=(1/1.5, 1.5),
cleaning_sensitivity_alternatives=(.25, .75),
clean_pruning_sensitivity_alternatives=(1/1.5, 1.5),
forward_fill_alternatives=(True, False),
verbose=False,
**kwargs):
Expand Down Expand Up @@ -2314,11 +2309,11 @@ def soiling_cods(energy_normalized_daily,
order_alternatives : tuple of tuples, default (('SR', 'SC', 'Rd'), ('SC', 'SR', 'Rd'))
Component estimation orders that will be tested during initial
model fitting.
detection_tuner_alternatives : tuple, default (.25, .75)
cleaning_sensitivity_alternatives : tuple, default (.25, .75)
Detection tuner values that will be tested during initial fitting.
Length must be >= 1. First and last values define limits of values
that will be used during bootstrapping.
pruning_tuner_alternatives : tuple, default (1/1.5, 1.5)
clean_pruning_sensitivity_alternatives : tuple, default (1/1.5, 1.5)
Pruning tuner values that will be tested during initial fitting.
Length must be >= 1. First and last values define limits of values
that will be used during bootstrapping.
Expand Down Expand Up @@ -2409,8 +2404,8 @@ def soiling_cods(energy_normalized_daily,
degradation_method=degradation_method,
process_noise=process_noise,
order_alternatives=order_alternatives,
detection_tuner_alternatives=detection_tuner_alternatives,
pruning_tuner_alternatives=pruning_tuner_alternatives,
cleaning_sensitivity_alternatives=cleaning_sensitivity_alternatives,
clean_pruning_sensitivity_alternatives=clean_pruning_sensitivity_alternatives,
forward_fill_alternatives=forward_fill_alternatives,
**kwargs)

Expand Down