from pyPhenology import models, utils
import numpy as np
import pytest
from copy import deepcopy
obs, predictors = utils.load_test_data(name='vaccinium', phenophase='budburst')
###########################################
# Setup various models
###########################################
########
# This is an ensemble of bootstrap models
alternating_bootstrap = models.BootstrapModel(core_model=models.Alternating,
num_bootstraps=10)
thermaltime_bootstrap = models.BootstrapModel(core_model=models.ThermalTime,
parameters={'t1':1,'T':0},
num_bootstraps=10)
uniforc_bootstrap = models.BootstrapModel(core_model=models.Uniforc,
parameters={'t1':1,},
num_bootstraps=10)
ensemble_bootstrap_model = models.Ensemble(core_models=[alternating_bootstrap,
thermaltime_bootstrap,
uniforc_bootstrap])
ensemble_bootstrap_model.fit(obs, predictors, optimizer_params='testing')
##########
# A simple ensemble model
tt1 = models.ThermalTime(parameters={'T':0})
tt2 = models.ThermalTime(parameters={'T':5})
uniforc = models.Uniforc(parameters={'t1':1})
from pyPhenology import utils
observations, temp = utils.load_test_data(name='vaccinium')
import pandas as pd
import time
models_to_use = ['ThermalTime', 'Alternating', 'Uniforc', 'MSB', 'Unichill']
methods_to_use = ['DE', 'BF']
sensible_defaults = ['testing', 'practical', 'intensive']
timings = []
for m in models_to_use:
for optimizer_method in methods_to_use:
for s_default in sensible_defaults:
model = utils.load_model(m)()
start_time = time.time()
model.fit(observations,
temp,
method=optimizer_method,
optimizer_params=s_default)
fitting_time = round(time.time() - start_time, 2)
num_parameters = len(model.get_params())
timings.append({
'model': m,
'num_parameters': num_parameters,
'method': optimizer_method,
'sensible_default': s_default,
'time': fitting_time
from pyPhenology import utils
import pytest
import numpy as np
obs, predictors = utils.load_test_data()
core_model_names = [
'Uniforc', 'Unichill', 'ThermalTime', 'Alternating', 'MSB', 'Linear',
'Sequential', 'M1', 'Naive', 'FallCooling'
]
# Setup a list of (model_name, fitted_model object)
fitted_models = [
utils.load_model(model_name)() for model_name in core_model_names
]
[
model.fit(obs, predictors, optimizer_params='testing')
for model in fitted_models
]
model_test_cases = list(zip(core_model_names, fitted_models))
#########################################################
@pytest.mark.parametrize('model_name, fitted_model', model_test_cases)
def test_predict_output_length(model_name, fitted_model):
"""Predict output length should equal input length"""
assert len(fitted_model.predict()) == len(obs)
def test_unknown_dataset_name(dataset):
with pytest.raises(ValueError):
obs, predictors = utils.load_test_data(name = dataset)
from pyPhenology import utils, models
from warnings import warn
import sys
import pytest
# Make sure some known parameters are estimated correctly
# The number roughly match the estimates from the original model
# codebase in github.com/sdtaylor/phenology_dataset_study.
# They don't match exactly because that original stuff uses DE
# while I'm using BF here for testing sake.
##############################
# Use a specific species dataset that should never change
leaves_obs, vaccinium_temp = utils.load_test_data(name='vaccinium',
phenophase=371)
flowers_obs, vaccinium_temp = utils.load_test_data(name='vaccinium',
phenophase=501)
aspen_leaves, aspen_temp = utils.load_test_data(name='aspen', phenophase=371)
# thorough but still relatively quick
thorough_DE_optimization = {
'method': 'DE',
'debug': True,
'optimizer_params': {
'seed': 1,
'popsize': 20,
'maxiter': 100,
'mutation': 1.5,
'recombination': 0.25
}
def test_invalid_dataset_phenophase(phenophase):
with pytest.raises(TypeError):
obs, predictors = utils.load_test_data(name='aspen', phenophase=phenophase)
def test_dataset_names(dataset):
obs, predictors = utils.load_test_data(dataset)
assert len(obs) > 0
def test_unknown_dataset_phenophase(phenophase):
with pytest.raises(ValueError):
obs, predictors = utils.load_test_data(name='aspen', phenophase=phenophase)
def test_dataset_phenophase(phenophase):
obs, predictors = utils.load_test_data(name='aspen', phenophase=phenophase)
assert len(obs) > 0
def test_invalid_dataset_name(dataset):
'dataset name needs to be a string'
with pytest.raises(TypeError):
obs, predictors = utils.load_test_data(name = dataset)
from pyPhenology import utils, models
from pyPhenology.models import validation
import pytest
import sys
obs, temp = utils.load_test_data()
model_test_cases = []
# Bootstrap requires some special arguments
model_test_cases.append({
'model_name': 'BootstrapModel',
'model_func': models.BootstrapModel,
'fit_params': {
'optimizer_params': 'testing'
},
'initial_params': {
'num_bootstraps': 10,
'core_model': models.ThermalTime
}
})
# The rest can all be tested the same way
model_names = [
'Uniforc', 'Unichill', 'ThermalTime', 'Alternating', 'MSB', 'Linear',
'Sequential'
]
for name in model_names:
model_test_cases.append({
'model_name': name,
'model_func': utils.load_model(name),
parameters={
't1': 1,
'T': 5
})
models_to_fit = {
'TT Temp 0': bootstrapped_tt_model_1,
'TT Temp 5': bootstrapped_tt_model_2
}
results = pd.DataFrame()
for dataset in datasets_to_use:
for phenophase in phenophases_to_use:
observations, predictors = utils.load_test_data(name=dataset,
phenophase=phenophase)
# Setup 20% train test split using pandas methods
observations_test = observations.sample(frac=0.2, random_state=1)
observations_train = observations[~observations.index.
isin(observations_test.index)]
observed_doy = observations_test.doy.values
for model_name, model in models_to_fit.items():
model.fit(observations_train,
predictors,
optimizer_params='practical')
# Using aggregation='none' in BoostrapModel predict
# returns results for all bootstrapped models in an
