Coverage for biobb_ml/neural_networks/classification_neural_network.py: 85%
192 statements
« prev ^ index » next coverage.py v7.6.1, created at 2024-10-03 14:57 +0000
1#!/usr/bin/env python3
3"""Module containing the ClassificationNeuralNetwork class and the command line interface."""
4import argparse
5import h5py
6import json
7import numpy as np
8import pandas as pd
9from biobb_common.generic.biobb_object import BiobbObject
10from tensorflow.python.keras.saving import hdf5_format
11from sklearn.preprocessing import scale
12from sklearn.model_selection import train_test_split
13from tensorflow.keras import Sequential
14from tensorflow.keras.layers import Dense
15from tensorflow.keras.callbacks import EarlyStopping
16from tensorflow import math
17from biobb_common.configuration import settings
18from biobb_common.tools import file_utils as fu
19from biobb_common.tools.file_utils import launchlogger
20from biobb_ml.neural_networks.common import check_input_path, check_output_path, getHeader, getTargetValue, getFeatures, getIndependentVarsList, getWeight, plotResultsClassMultCM, plotResultsClassBinCM
23class ClassificationNeuralNetwork(BiobbObject):
24 """
25 | biobb_ml ClassificationNeuralNetwork
26 | Wrapper of the TensorFlow Keras Sequential method for classification.
27 | Trains and tests a given dataset and save the complete model for a Neural Network Classification. Visit the `Sequential documentation page <https://www.tensorflow.org/api_docs/python/tf/keras/Sequential>`_ in the TensorFlow Keras official website for further information.
29 Args:
30 input_dataset_path (str): Path to the input dataset. File type: input. `Sample file <https://github.com/bioexcel/biobb_ml/raw/master/biobb_ml/test/data/neural_networks/dataset_classification.csv>`_. Accepted formats: csv (edam:format_3752).
31 output_model_path (str): Path to the output model file. File type: output. `Sample file <https://github.com/bioexcel/biobb_ml/raw/master/biobb_ml/test/reference/neural_networks/ref_output_model_classification.h5>`_. Accepted formats: h5 (edam:format_3590).
32 output_test_table_path (str) (Optional): Path to the test table file. File type: output. `Sample file <https://github.com/bioexcel/biobb_ml/raw/master/biobb_ml/test/reference/neural_networks/ref_output_test_classification.csv>`_. Accepted formats: csv (edam:format_3752).
33 output_plot_path (str) (Optional): Loss, accuracy and MSE plots. File type: output. `Sample file <https://github.com/bioexcel/biobb_ml/raw/master/biobb_ml/test/reference/neural_networks/ref_output_plot_classification.png>`_. Accepted formats: png (edam:format_3603).
34 properties (dic - Python dictionary object containing the tool parameters, not input/output files):
35 * **features** (*dict*) - ({}) Independent variables or columns from your dataset you want to train. You can specify either a list of columns names from your input dataset, a list of columns indexes or a range of columns indexes. Formats: { "columns": ["column1", "column2"] } or { "indexes": [0, 2, 3, 10, 11, 17] } or { "range": [[0, 20], [50, 102]] }. In case of mulitple formats, the first one will be picked.
36 * **target** (*dict*) - ({}) Dependent variable you want to predict from your dataset. You can specify either a column name or a column index. Formats: { "column": "column3" } or { "index": 21 }. In case of mulitple formats, the first one will be picked.
37 * **weight** (*dict*) - ({}) Weight variable from your dataset. You can specify either a column name or a column index. Formats: { "column": "column3" } or { "index": 21 }. In case of multiple formats, the first one will be picked.
38 * **validation_size** (*float*) - (0.2) [0~1|0.05] Represents the proportion of the dataset to include in the validation split. It should be between 0.0 and 1.0.
39 * **test_size** (*float*) - (0.1) [0~1|0.05] Represents the proportion of the dataset to include in the test split. It should be between 0.0 and 1.0.
40 * **hidden_layers** (*list*) - (None) List of dictionaries with hidden layers values. Format: [ { 'size': 50, 'activation': 'relu' } ].
41 * **output_layer_activation** (*string*) - ("softmax") Activation function to use in the output layer. Values: sigmoid (Sigmoid activation function: sigmoid[x] = 1 / [1 + exp[-x]]), tanh (Hyperbolic tangent activation function), relu (Applies the rectified linear unit activation function), softmax (Softmax converts a real vector to a vector of categorical probabilities).
42 * **optimizer** (*string*) - ("Adam") Name of optimizer instance. Values: Adadelta (Adadelta optimization is a stochastic gradient descent method that is based on adaptive learning rate per dimension to address two drawbacks: the continual decay of learning rates throughout training and the need for a manually selected global learning rate), Adagrad (Adagrad is an optimizer with parameter-specific learning rates; which are adapted relative to how frequently a parameter gets updated during training. The more updates a parameter receives; the smaller the updates), Adam (Adam optimization is a stochastic gradient descent method that is based on adaptive estimation of first-order and second-order moments), Adamax (It is a variant of Adam based on the infinity norm. Default parameters follow those provided in the paper. Adamax is sometimes superior to adam; specially in models with embeddings), Ftrl (Optimizer that implements the FTRL algorithm), Nadam (Much like Adam is essentially RMSprop with momentum; Nadam is Adam with Nesterov momentum), RMSprop (Optimizer that implements the RMSprop algorithm), SGD (Gradient descent -with momentum- optimizer).
43 * **learning_rate** (*float*) - (0.02) [0~100|0.01] Determines the step size at each iteration while moving toward a minimum of a loss function
44 * **batch_size** (*int*) - (100) [0~1000|1] Number of samples per gradient update.
45 * **max_epochs** (*int*) - (100) [0~1000|1] Number of epochs to train the model. As the early stopping is enabled, this is a maximum.
46 * **normalize_cm** (*bool*) - (False) Whether or not to normalize the confusion matrix.
47 * **random_state** (*int*) - (5) [1~1000|1] Controls the shuffling applied to the data before applying the split. .
48 * **scale** (*bool*) - (False) Whether or not to scale the input dataset.
49 * **remove_tmp** (*bool*) - (True) [WF property] Remove temporal files.
50 * **restart** (*bool*) - (False) [WF property] Do not execute if output files exist.
51 * **sandbox_path** (*str*) - ("./") [WF property] Parent path to the sandbox directory.
53 Examples:
54 This is a use example of how to use the building block from Python::
56 from biobb_ml.neural_networks.classification_neural_network import classification_neural_network
57 prop = {
58 'features': {
59 'columns': [ 'column1', 'column2', 'column3' ]
60 },
61 'target': {
62 'column': 'target'
63 },
64 'validation_size': 0.2,
65 'test_size': .33,
66 'hidden_layers': [
67 {
68 'size': 10,
69 'activation': 'relu'
70 },
71 {
72 'size': 8,
73 'activation': 'relu'
74 }
75 ],
76 'optimizer': 'Adam',
77 'learning_rate': 0.01,
78 'batch_size': 32,
79 'max_epochs': 150
80 }
81 classification_neural_network(input_dataset_path='/path/to/myDataset.csv',
82 output_model_path='/path/to/newModel.h5',
83 output_test_table_path='/path/to/newTable.csv',
84 output_plot_path='/path/to/newPlot.png',
85 properties=prop)
87 Info:
88 * wrapped_software:
89 * name: TensorFlow Keras Sequential
90 * version: >2.1.0
91 * license: MIT
92 * ontology:
93 * name: EDAM
94 * schema: http://edamontology.org/EDAM.owl
96 """
98 def __init__(self, input_dataset_path, output_model_path,
99 output_test_table_path=None, output_plot_path=None, properties=None, **kwargs) -> None:
100 properties = properties or {}
102 # Call parent class constructor
103 super().__init__(properties)
104 self.locals_var_dict = locals().copy()
106 # Input/Output files
107 self.io_dict = {
108 "in": {"input_dataset_path": input_dataset_path},
109 "out": {"output_model_path": output_model_path, "output_test_table_path": output_test_table_path, "output_plot_path": output_plot_path}
110 }
112 # Properties specific for BB
113 self.features = properties.get('features', {})
114 self.target = properties.get('target', {})
115 self.weight = properties.get('weight', {})
116 self.validation_size = properties.get('validation_size', 0.1)
117 self.test_size = properties.get('test_size', 0.1)
118 self.hidden_layers = properties.get('hidden_layers', [])
119 self.output_layer_activation = properties.get('output_layer_activation', 'softmax')
120 self.optimizer = properties.get('optimizer', 'Adam')
121 self.learning_rate = properties.get('learning_rate', 0.02)
122 self.batch_size = properties.get('batch_size', 100)
123 self.max_epochs = properties.get('max_epochs', 100)
124 self.normalize_cm = properties.get('normalize_cm', False)
125 self.random_state = properties.get('random_state', 5)
126 self.scale = properties.get('scale', False)
127 self.properties = properties
129 # Check the properties
130 self.check_properties(properties)
131 self.check_arguments()
133 def check_data_params(self, out_log, err_log):
134 """ Checks all the input/output paths and parameters """
135 self.io_dict["in"]["input_dataset_path"] = check_input_path(self.io_dict["in"]["input_dataset_path"], "input_dataset_path", False, out_log, self.__class__.__name__)
136 self.io_dict["out"]["output_model_path"] = check_output_path(self.io_dict["out"]["output_model_path"], "output_model_path", False, out_log, self.__class__.__name__)
137 self.io_dict["out"]["output_test_table_path"] = check_output_path(self.io_dict["out"]["output_test_table_path"], "output_test_table_path", True, out_log, self.__class__.__name__)
138 self.io_dict["out"]["output_plot_path"] = check_output_path(self.io_dict["out"]["output_plot_path"], "output_plot_path", True, out_log, self.__class__.__name__)
140 def build_model(self, input_shape, output_size):
141 """ Builds Neural network according to hidden_layers property """
143 # create model
144 model = Sequential([])
146 # if no hidden_layers provided, create manually a hidden layer with default values
147 if not self.hidden_layers:
148 self.hidden_layers = [{'size': 50, 'activation': 'relu'}]
150 # generate hidden_layers
151 for i, layer in enumerate(self.hidden_layers):
152 if i == 0:
153 model.add(Dense(layer['size'], activation=layer['activation'], kernel_initializer='he_normal', input_shape=input_shape)) # 1st hidden layer
154 else:
155 model.add(Dense(layer['size'], activation=layer['activation'], kernel_initializer='he_normal'))
157 model.add(Dense(output_size, activation=self.output_layer_activation)) # output layer
159 return model
161 @launchlogger
162 def launch(self) -> int:
163 """Execute the :class:`ClassificationNeuralNetwork <neural_networks.classification_neural_network.ClassificationNeuralNetwork>` neural_networks.classification_neural_network.ClassificationNeuralNetwork object."""
165 # check input/output paths and parameters
166 self.check_data_params(self.out_log, self.err_log)
168 # Setup Biobb
169 if self.check_restart():
170 return 0
171 self.stage_files()
173 # load dataset
174 fu.log('Getting dataset from %s' % self.io_dict["in"]["input_dataset_path"], self.out_log, self.global_log)
175 if 'columns' in self.features:
176 labels = getHeader(self.io_dict["in"]["input_dataset_path"])
177 skiprows = 1
178 else:
179 labels = None
180 skiprows = None
181 data = pd.read_csv(self.io_dict["in"]["input_dataset_path"], header=None, sep="\\s+|;|:|,|\t", engine="python", skiprows=skiprows, names=labels)
183 targets_list = data[getTargetValue(self.target)].to_numpy()
185 X = getFeatures(self.features, data, self.out_log, self.__class__.__name__)
186 fu.log('Features: [%s]' % (getIndependentVarsList(self.features)), self.out_log, self.global_log)
187 # target
188 # y = getTarget(self.target, data, self.out_log, self.__class__.__name__)
189 fu.log('Target: %s' % (str(getTargetValue(self.target))), self.out_log, self.global_log)
190 # weights
191 if self.weight:
192 w = getWeight(self.weight, data, self.out_log, self.__class__.__name__)
194 # shuffle dataset
195 fu.log('Shuffling dataset', self.out_log, self.global_log)
196 shuffled_indices = np.arange(X.shape[0])
197 np.random.shuffle(shuffled_indices)
198 np_X = X.to_numpy()
199 shuffled_X = np_X[shuffled_indices]
200 shuffled_y = targets_list[shuffled_indices]
201 if self.weight:
202 shuffled_w = w[shuffled_indices]
204 # train / test split
205 fu.log('Creating train and test sets', self.out_log, self.global_log)
206 arrays_sets = (shuffled_X, shuffled_y)
207 # if user provide weights
208 if self.weight:
209 arrays_sets = arrays_sets + (shuffled_w,)
210 X_train, X_test, y_train, y_test, w_train, w_test = train_test_split(*arrays_sets, test_size=self.test_size, random_state=self.random_state)
211 else:
212 X_train, X_test, y_train, y_test = train_test_split(*arrays_sets, test_size=self.test_size, random_state=self.random_state)
214 # scale dataset
215 if self.scale:
216 fu.log('Scaling dataset', self.out_log, self.global_log)
217 X_train = scale(X_train)
219 # build model
220 fu.log('Building model', self.out_log, self.global_log)
221 model = self.build_model((X_train.shape[1],), np.unique(y_train).size)
223 # model summary
224 stringlist = []
225 model.summary(print_fn=lambda x: stringlist.append(x))
226 model_summary = "\n".join(stringlist)
227 fu.log('Model summary:\n\n%s\n' % model_summary, self.out_log, self.global_log)
229 # get optimizer
230 mod = __import__('tensorflow.keras.optimizers', fromlist=[self.optimizer])
231 opt_class = getattr(mod, self.optimizer)
232 opt = opt_class(lr=self.learning_rate)
233 # compile model
234 model.compile(optimizer=opt, loss='sparse_categorical_crossentropy', metrics=['accuracy', 'mse'])
236 # fitting
237 fu.log('Training model', self.out_log, self.global_log)
238 # set an early stopping mechanism
239 # set patience=2, to be a bit tolerant against random validation loss increases
240 early_stopping = EarlyStopping(patience=2)
242 if self.weight:
243 sample_weight = w_train
244 class_weight = []
245 else:
246 # TODO: class_weight not working since TF 2.4.1 update
247 # fu.log('No weight provided, class_weight will be estimated from the target data', self.out_log, self.global_log)
248 fu.log('No weight provided', self.out_log, self.global_log)
249 sample_weight = None
250 class_weight = [] # compute_class_weight('balanced', np.unique(y_train), y_train)
252 print(class_weight)
253 # fit the model
254 mf = model.fit(X_train,
255 y_train,
256 class_weight=class_weight,
257 sample_weight=sample_weight,
258 batch_size=self.batch_size,
259 epochs=self.max_epochs,
260 callbacks=[early_stopping],
261 validation_split=self.validation_size,
262 verbose=1)
264 fu.log('Total epochs performed: %s' % len(mf.history['loss']), self.out_log, self.global_log)
266 train_metrics = pd.DataFrame()
267 train_metrics['metric'] = ['Train loss', ' Train accuracy', 'Train MSE', 'Validation loss', 'Validation accuracy', 'Validation MSE']
268 train_metrics['coefficient'] = [mf.history['loss'][-1], mf.history['accuracy'][-1], mf.history['mse'][-1], mf.history['val_loss'][-1], mf.history['val_accuracy'][-1], mf.history['val_mse'][-1]]
270 fu.log('Training metrics\n\nTRAINING METRICS TABLE\n\n%s\n' % train_metrics, self.out_log, self.global_log)
272 # confusion matrix
273 train_predictions = model.predict(X_train)
274 train_predictions = np.around(train_predictions, decimals=2)
275 norm_pred = []
276 [norm_pred.append(np.argmax(pred, axis=0)) for pred in train_predictions]
277 cnf_matrix_train = math.confusion_matrix(y_train, norm_pred).numpy()
278 np.set_printoptions(precision=2)
279 if self.normalize_cm:
280 cnf_matrix_train = cnf_matrix_train.astype('float') / cnf_matrix_train.sum(axis=1)[:, np.newaxis]
281 cm_type = 'NORMALIZED CONFUSION MATRIX'
282 else:
283 cm_type = 'CONFUSION MATRIX, WITHOUT NORMALIZATION'
285 fu.log('Calculating confusion matrix for training dataset\n\n%s\n\n%s\n' % (cm_type, cnf_matrix_train), self.out_log, self.global_log)
287 # testing
288 if self.scale:
289 X_test = scale(X_test)
290 fu.log('Testing model', self.out_log, self.global_log)
291 test_loss, test_accuracy, test_mse = model.evaluate(X_test, y_test)
293 test_metrics = pd.DataFrame()
294 test_metrics['metric'] = ['Test loss', ' Test accuracy', 'Test MSE']
295 test_metrics['coefficient'] = [test_loss, test_accuracy, test_mse]
297 fu.log('Testing metrics\n\nTESTING METRICS TABLE\n\n%s\n' % test_metrics, self.out_log, self.global_log)
299 # predict data from X_test
300 test_predictions = model.predict(X_test)
301 test_predictions = np.around(test_predictions, decimals=2)
302 tpr = tuple(map(tuple, test_predictions))
304 test_table = pd.DataFrame()
305 test_table['P' + np.array2string(np.unique(y_train))] = tpr
306 test_table['target'] = y_test
308 fu.log('TEST DATA\n\n%s\n' % test_table, self.out_log, self.global_log)
310 # confusion matrix
311 norm_pred = []
312 [norm_pred.append(np.argmax(pred, axis=0)) for pred in test_predictions]
313 cnf_matrix_test = math.confusion_matrix(y_test, norm_pred).numpy()
314 np.set_printoptions(precision=2)
315 if self.normalize_cm:
316 cnf_matrix_test = cnf_matrix_test.astype('float') / cnf_matrix_test.sum(axis=1)[:, np.newaxis]
317 cm_type = 'NORMALIZED CONFUSION MATRIX'
318 else:
319 cm_type = 'CONFUSION MATRIX, WITHOUT NORMALIZATION'
321 fu.log('Calculating confusion matrix for testing dataset\n\n%s\n\n%s\n' % (cm_type, cnf_matrix_test), self.out_log, self.global_log)
323 # save test data
324 if (self.io_dict["out"]["output_test_table_path"]):
325 fu.log('Saving testing data to %s' % self.io_dict["out"]["output_test_table_path"], self.out_log, self.global_log)
326 test_table.to_csv(self.io_dict["out"]["output_test_table_path"], index=False, header=True)
328 # create test plot
329 if (self.io_dict["out"]["output_plot_path"]):
330 vs = np.unique(targets_list)
331 vs.sort()
332 if len(vs) > 2:
333 plot = plotResultsClassMultCM(mf.history, cnf_matrix_train, cnf_matrix_test, self.normalize_cm, vs)
334 fu.log('Saving confusion matrix plot to %s' % self.io_dict["out"]["output_plot_path"], self.out_log, self.global_log)
335 else:
336 plot = plotResultsClassBinCM(mf.history, train_predictions, test_predictions, y_train, y_test, cnf_matrix_train, cnf_matrix_test, self.normalize_cm, vs)
337 fu.log('Saving binary classifier evaluator plot to %s' % self.io_dict["out"]["output_plot_path"], self.out_log, self.global_log)
338 plot.savefig(self.io_dict["out"]["output_plot_path"], dpi=150)
340 # save model and parameters
341 vs = np.unique(targets_list)
342 vs.sort()
343 vars_obj = {
344 'features': self.features,
345 'target': self.target,
346 'scale': self.scale,
347 'vs': vs.tolist(),
348 'type': 'classification'
349 }
350 variables = json.dumps(vars_obj)
351 fu.log('Saving model to %s' % self.io_dict["out"]["output_model_path"], self.out_log, self.global_log)
352 with h5py.File(self.io_dict["out"]["output_model_path"], mode='w') as f:
353 hdf5_format.save_model_to_hdf5(model, f)
354 f.attrs['variables'] = variables
356 # Copy files to host
357 self.copy_to_host()
359 self.tmp_files.extend([
360 self.stage_io_dict.get("unique_dir")
361 ])
362 self.remove_tmp_files()
364 self.check_arguments(output_files_created=True, raise_exception=False)
366 return 0
369def classification_neural_network(input_dataset_path: str, output_model_path: str, output_test_table_path: str = None, output_plot_path: str = None, properties: dict = None, **kwargs) -> int:
370 """Execute the :class:`AutoencoderNeuralNetwork <neural_networks.classification_neural_network.AutoencoderNeuralNetwork>` class and
371 execute the :meth:`launch() <neural_networks.classification_neural_network.AutoencoderNeuralNetwork.launch>` method."""
373 return ClassificationNeuralNetwork(input_dataset_path=input_dataset_path,
374 output_model_path=output_model_path,
375 output_test_table_path=output_test_table_path,
376 output_plot_path=output_plot_path,
377 properties=properties, **kwargs).launch()
380def main():
381 """Command line execution of this building block. Please check the command line documentation."""
382 parser = argparse.ArgumentParser(description="Wrapper of the TensorFlow Keras Sequential method.", formatter_class=lambda prog: argparse.RawTextHelpFormatter(prog, width=99999))
383 parser.add_argument('--config', required=False, help='Configuration file')
385 # Specific args of each building block
386 required_args = parser.add_argument_group('required arguments')
387 required_args.add_argument('--input_dataset_path', required=True, help='Path to the input dataset. Accepted formats: csv.')
388 required_args.add_argument('--output_model_path', required=True, help='Path to the output model file. Accepted formats: h5.')
389 parser.add_argument('--output_test_table_path', required=False, help='Path to the test table file. Accepted formats: csv.')
390 parser.add_argument('--output_plot_path', required=False, help='Loss, accuracy and MSE plots. Accepted formats: png.')
392 args = parser.parse_args()
393 args.config = args.config or "{}"
394 properties = settings.ConfReader(config=args.config).get_prop_dic()
396 # Specific call of each building block
397 classification_neural_network(input_dataset_path=args.input_dataset_path,
398 output_model_path=args.output_model_path,
399 output_test_table_path=args.output_test_table_path,
400 output_plot_path=args.output_plot_path,
401 properties=properties)
404if __name__ == '__main__':
405 main()