objective_thresholds is a Python package for objective threshold level calculation based on a flow duration curve (FDC). This package contains tools for threshold calculation for low flow and flood analysis, contained in the accompanying lowflow and flood modules. The functions included in the modules allow for the computation of a single threshold value (function threshold) on a given lower (higher) range of the FDC, or multiple (function multiple) thresholds if additional division is needed (i.e., “shallow” and “deep” low flows). Further implementations and modules will be added in the future to allow for additional threshold calculations and analysis at full range of FDC.
If you use objective_thresholds in a scientific publication, please include the reference below:
Raczyński K., Dyer J., 2022, Development of an Objective Low Flow Identification Method Using Breakpoint Analysis, Water, 14(14), 2212, https://doi.org/10.3390/w14142212
Official repository website address: https://github.com/chrisrac/objective_thresholds/
Required Jenkspy library. Depending on the version of Jenkspy available, use the appropriate release:
- if Jenkspy v.0.2.4 or older is used, install it with pip: pip install objective_thresholds or download release v.0.1.
- if Jenkspy 0.3 or newer is used, download release v.0.2. The main difference is in Jenkspy 0.3 and newer, where the argument "nb_class" was renamed to "n_classes". This naming affects the way this package works (by returning a TypeError if the wrong version is used). The difference between Objective Thresholds releases 0.1 and 0.2 is in the argument naming only. These are meant to be used with varying Jenkspy versions.
The objective_thresholds package requires the following:
- Python 3.4+
- Pandas
- NumPy
- Jenkspy v.0.2.4 (version 0.3+ is not supported)
The objective_thresholds package currently includes the modules lowflow and flood with functions available in each. To use the package, once installed it should be imported to Python: import objective_thresholds.lowflow as olf or: import objective_thresholds.flood as ofd for flood studies. The function threshold allows for the computation of low flow (flood) thresholds using breakpoints through the application of the Fisher-Jenks algorithm, which is applied to the lower (higher) range of the FDC (calculated from provided dataset: input_data). By default, the lowest (highest) 35% of the FDC is used for low flow (flood) threshold calculations, and one threshold value is returned. The function accepts as input any array-like data, including: lists, pandas Series, numpy array, or slices of pandas DataFrame. Optionally, the limit of the lower (higher) FDC can be set by the user if a value other than the default 35th percentile is needed. If the input data are highly discretized (contains numerous repetitive values), the method argument controls the result: • leave – breaks the computation, Exception is returned; • max – the maximum value in the lower FDC range is returned; • median – the median value in the lower FDC range is returned. The threshold function returns one float or int type value representing the identified threshold, depending on the input data type. Some examples of usage:
# importing lowflow module from objective_thresholds package
>>> import objective_thresholds.lowflow as olf
>>> import pandas as pd
# importing example data from testdata.csv file (available at source repository)
>>> data = pd.read_csv('home/testdata.csv')
# checking data structure
>>> data.head()
id flow1 flow2
0 0 2.71 2.84
1 1 0.31 0.34
2 2 0.16 0.18
3 3 0.23 0.26
4 4 0.26 0.28
# calculating threshold value for first data series using default settings (35th percentile; limit=0.35)
>>> threshold_value = olf.threshold(data['flow1'])
# results in: 0.13
# modifying range of lower FDC to 50th percentile
>>> threshold_50p = olf.threshold(data['flow1'], limit=0.5)
# results in: 0.19
# an Exception is raised if fewer than 5 unique values are present
>>> threshold_10p = olf.threshold(data['flow2'], limit=0.10)
Exception: Unique values of lower FDC range are less than 5
and used method is "leave", no thresholds are returned.
# controlling output by changing the method for discretized data
threshold_max = olf.threshold(data['flow2'], limit=0.10, method = 'max')
# results in: 0.11The function multiple allows for the computation of multiple low flow (flood) thresholds using breakpoints through the application of the Fisher-Jenks algorithm, which is applied to the lower (higher) range of the FDC (provided as input_data). By default, the lowest (highest) 35% of the FDC is used for low flow (flood) threshold calculations, and multiple threshold values are returned. This function is to be used when at least two thresholds are needed, such as when dividing the lower FDC range into “shallow” and “deep” low flow events. The function accepts as input any array-like data, including: lists, pandas Series, numpy array, or slices of pandas DataFrame. Argument breaks controls the number of thresholds to be returned. Optionally, the limit of the lower (higher) FDC can be set by the user if a value other than the default 35th percentile is needed.. If the input data are highly discretized (contains numerous repetitive values), the method argument controls the result: • leave – breaks the computation, Exception is returned; • max – the maximum value in the lower FDC range is returned; • median – the median value in the lower FDC range is returned. Note that the implementation of the max or median arguments causes the multiple function to return a single value, regardless of the number of breakpoints requested. The multiple function returns a list of float or int type values, depending on the input data type, representing the identified thresholds. Some examples of usage:
# importing lowflow module from objective_thresholds package
>>> import objective_thresholds.lowflow as olf
>>> import pandas as pd
# importing example data from testdata.csv file (available at source repository)
>>> data = pd.read_csv('home/testdata.csv')
# checking data structure
>>> data.head()
id flow1 flow2
0 0 2.71 2.84
1 1 0.31 0.34
2 2 0.16 0.18
3 3 0.23 0.26
4 4 0.26 0.28
# calculating threshold value for first data series using default settings (limit=0.35; breaks=2)
>>> threshold_value = olf.multiple(data['flow1'])
# results in: [0.1, 0.16]
# modifying range of lower FDC to 50th percentile
>>> threshold_50p = olf.multiple(data['flow1'], limit=0.5)
# results in: [0.14, 0.24]
# modifying the number of thresholds to be returned
>>> threshold_3th = olf.multiple(data['flow1'], breaks=3)
# results in: [0.09, 0.13, 0.17]
# an Exception is raised if fewer than 5 unique values are present
>>> threshold_10p = olf.multiple(data['flow2'], limit=0.10)
Exception: Unique values of lower FDC range are less than 5
and used method is "leave", no thresholds are returned.
# controlling output by changing the method for discretized data
threshold_max = olf.multiple(data['flow2'], limit=0.10, method = 'median', breaks=6)
# results in: 0.08 as only one value is returned by max or median method, regardless of breaks settingCopyright (c) 2022, Raczynski K., Dyer J.
Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions: The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.