tsadmetrics 0.1.4__py3-none-any.whl → 0.1.5__py3-none-any.whl

tsadmetrics/_tsadeval/prts/time_series_metrics/fscore.py

@@ -0,0 +1,61 @@
+import numpy as np
+
+from ..base.time_series_metrics import BaseTimeSeriesMetrics
+from .precision import TimeSeriesPrecision
+from .recall import TimeSeriesRecall
+
+
+class TimeSeriesFScore(BaseTimeSeriesMetrics):
+    """ This class calculates f-score for time series"""
+
+    def __init__(self, beta=1.0, p_alpha=0.0, r_alpha=0.0, cardinality="one", p_bias="flat", r_bias="flat"):
+        """Constructor
+
+        Args:
+            beta (float, optional): determines the weight of recall in the combined score.. Defaults to 1.0.
+            p_alpha (float, optional): alpha of precision, 0<=alpha_p<=1. Defaults to 0.0.
+            r_alpha (float, optional): alpha of recall, 0<=alpha<=1. Defaults to 0.0.
+            cardinality (str, optional): ["one", "reciprocal", "udf_gamma"]. Defaults to "one".
+            p_bias (str, optional): bias of precision, ["flat", "front", "middle", "back"]. Defaults to "flat".
+            r_bias (str, optional): bias of recall, ["flat", "front", "middle", "back"]. Defaults to "flat".
+        """
+
+        assert beta >= 0
+
+        self.beta = beta
+        self.p_alpha = p_alpha
+        self.r_alpha = r_alpha
+        self.cardinality = cardinality
+        self.p_bias = p_bias
+        self.r_bias = r_bias
+
+    def score(self, real: np.ndarray, pred: np.ndarray) -> float:
+        """Computing fbeta score
+
+        Args:
+            real (np.ndarray):
+                One-dimensional array of correct answers with values of 1 or 0.
+            pred (np.ndarray):
+                One-dimensional array of predicted answers with values of 1 or 0.
+
+        Returns:
+            float: fbeta
+        """
+
+        assert isinstance(real, np.ndarray) or isinstance(real, list)
+        assert isinstance(pred, np.ndarray) or isinstance(pred, list)
+
+        if not isinstance(real, np.ndarray):
+            real = np.array(real)
+        if not isinstance(pred, np.ndarray):
+            pred = np.array(pred)
+
+        precision = TimeSeriesPrecision(self.p_alpha, self.cardinality, self.p_bias).score(real, pred)
+        recall = TimeSeriesRecall(self.r_alpha, self.cardinality, self.r_bias).score(real, pred)
+
+        if precision + recall != 0:
+            f_beta = (1 + self.beta ** 2) * precision * recall / (self.beta ** 2 * precision + recall)
+        else:
+            f_beta = 0
+
+        return f_beta

tsadmetrics/_tsadeval/prts/time_series_metrics/precision.py

@@ -0,0 +1,86 @@
+from typing import Union
+
+import numpy as np
+
+from ..base.time_series_metrics import BaseTimeSeriesMetrics
+
+
+class TimeSeriesPrecision(BaseTimeSeriesMetrics):
+    """ This class calculates precision for time series"""
+
+    def __init__(self, alpha=0.0, cardinality="one", bias="flat"):
+        """Constructor
+
+        Args:
+            alpha (float, optional): 0 <= alpha <= 1. Defaults to 0.0.
+            cardinality (str, optional): ["one", "reciprocal", "udf_gamma"]. Defaults to "one".
+            bias (str, optional): ["flat", "front", "middle", "back"]. Defaults to "flat".
+        """
+
+        assert (alpha >= 0) & (alpha <= 1)
+        assert cardinality in ["one", "reciprocal", "udf_gamma"]
+        assert bias in ["flat", "front", "middle", "back"]
+
+        self.alpha = alpha
+        self.cardinality = cardinality
+        self.bias = bias
+
+    def score(self, real: Union[np.ndarray, list], pred: Union[np.ndarray, list]) -> float:
+        """Computing precision score
+
+        Args:
+            real (np.ndarray or list):
+                One-dimensional array of correct answers with values of 1 or 0.
+            pred (np.ndarray or list):
+                One-dimensional array of predicted answers with values of 1 or 0.
+
+        Returns:
+            float: precision
+        """
+
+        assert isinstance(real, np.ndarray) or isinstance(real, list)
+        assert isinstance(pred, np.ndarray) or isinstance(pred, list)
+
+        if not isinstance(real, np.ndarray):
+            real = np.array(real)
+        if not isinstance(pred, np.ndarray):
+            pred = np.array(pred)
+
+        real_anomalies, predicted_anomalies = self._prepare_data(real, pred)
+        precision = self._update_precision(real_anomalies, predicted_anomalies)
+
+        return precision
+
+    def _update_precision(self, real_anomalies: np.ndarray, predicted_anomalies: np.ndarray) -> float:
+        """Update precision
+
+        Args:
+            real_anomalies (np.ndarray):
+                2-dimensional array of the first and last indexes of each real anomaly range.
+                e.g. np.array([[1933,  1953],[1958,  2000], ...])
+            predicted_anomalies (np.ndarray):
+                2-dimensional array of the first and last indexes of each predicted anomaly range.
+                e.g. np.array([[1933,  1953],[1958,  2000], ...])
+
+        Returns:
+            float: precision
+        """
+        precision = 0
+        if len(predicted_anomalies) == 0:
+            return 0
+        for i in range(len(predicted_anomalies)):
+            range_p = predicted_anomalies[i, :]
+            omega_reward = 0
+            overlap_count = [0]
+            for j in range(len(real_anomalies)):
+                range_r = real_anomalies[j, :]
+                omega_reward += self._compute_omega_reward(range_p, range_r, overlap_count)
+            overlap_reward = self._gamma_function(overlap_count) * omega_reward
+            if overlap_count[0] > 0:
+                existence_reward = 1
+            else:
+                existence_reward = 0
+
+            precision += self.alpha * existence_reward + (1 - self.alpha) * overlap_reward
+        precision /= len(predicted_anomalies)
+        return precision

tsadmetrics/_tsadeval/prts/time_series_metrics/precision_recall.py

@@ -0,0 +1,21 @@
+from typing import Any
+
+import numpy as np
+
+from prts.base.time_series_metrics import BaseTimeSeriesMetrics
+
+
+class TimeSeriesPrecisionRecall(BaseTimeSeriesMetrics):
+    """ This class calculates precision and recall for time series """
+
+    def score(self, real: np.ndarray, pred: np.ndarray) -> Any:
+        """
+
+        Args:
+            real:
+            pred:
+
+        Returns:
+
+        """
+        # TODO: impl

tsadmetrics/_tsadeval/prts/time_series_metrics/recall.py

@@ -0,0 +1,85 @@
+from typing import Union
+
+import numpy as np
+
+from ..base.time_series_metrics import BaseTimeSeriesMetrics
+
+
+class TimeSeriesRecall(BaseTimeSeriesMetrics):
+    """ This class calculates recall for time series """
+
+    def __init__(self, alpha=0.0, cardinality="one", bias="flat"):
+        """Constructor
+
+        Args:
+            alpha (float, optional): 0 <= alpha <= 1. Defaults to 0.0.
+            cardinality (str, optional): ["one", "reciprocal", "udf_gamma"]. Defaults to "one".
+            bias (str, optional): ["flat", "front", "middle", "back"]. Defaults to "flat".
+        """
+
+        assert (alpha >= 0) & (alpha <= 1)
+        assert cardinality in ["one", "reciprocal", "udf_gamma"]
+        assert bias in ["flat", "front", "middle", "back"]
+
+        self.alpha = alpha
+        self.cardinality = cardinality
+        self.bias = bias
+
+    def score(self, real: Union[np.ndarray, list], pred: Union[np.ndarray, list]) -> float:
+        """Computing recall score
+
+        Args:
+            real (np.ndarray or list):
+                One-dimensional array of correct answers with values of 1 or 0.
+            pred (np.ndarray or list):
+                One-dimensional array of predicted answers with values of 1 or 0.
+        Returns:
+            float: recall
+        """
+
+        assert isinstance(real, np.ndarray) or isinstance(real, list)
+        assert isinstance(pred, np.ndarray) or isinstance(pred, list)
+
+        if not isinstance(real, np.ndarray):
+            real = np.array(real)
+        if not isinstance(pred, np.ndarray):
+            pred = np.array(pred)
+
+        real_anomalies, predicted_anomalies = self._prepare_data(real, pred)
+        recall = self._update_recall(real_anomalies, predicted_anomalies)
+
+        return recall
+
+    def _update_recall(self, real_anomalies: np.ndarray, predicted_anomalies: np.ndarray) -> float:
+        """Update recall
+        Args:
+            real_anomalies (np.ndarray):
+                2-dimensional array of the first and last indexes of each real anomaly range.
+                e.g. np.array([[1933,  1953],[1958,  2000], ...])
+            predicted_anomalies (np.ndarray):
+                2-dimensional array of the first and last indexes of each predicted anomaly range.
+                e.g. np.array([[1933,  1953],[1958,  2000], ...])
+        Returns:
+            float: recall
+        """
+
+        recall = 0
+        if len(real_anomalies) == 0:
+            return 0
+        for i in range(len(real_anomalies)):
+            omega_reward = 0
+            overlap_count = [0]
+            range_r = real_anomalies[i, :]
+            for j in range(len(predicted_anomalies)):
+                range_p = predicted_anomalies[j, :]
+                omega_reward += self._compute_omega_reward(range_r, range_p, overlap_count)
+            overlap_reward = self._gamma_function(overlap_count) * omega_reward
+
+            if overlap_count[0] > 0:
+                existence_reward = 1
+            else:
+                existence_reward = 0
+
+            recall += self.alpha * existence_reward + (1 - self.alpha) * overlap_reward
+        recall /= len(real_anomalies)
+        return recall

{tsadmetrics-0.1.4.dist-info → tsadmetrics-0.1.5.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: tsadmetrics
-Version: 0.1.4
+Version: 0.1.5
 Summary: Librería para evaluación de detección de anomalías en series temporales
 Home-page: https://github.com/pathsko/TSADmetrics
 Author: Pedro Rafael Velasco Priego

tsadmetrics-0.1.5.dist-info/RECORD

@@ -0,0 +1,62 @@
+entorno/bin/activate_this.py,sha256=45dnJsdtOWIt5LtVSBmBfB8E7AlKcnhnZe9e3WGclak,1199
+entorno/bin/rst2html.py,sha256=H9TATamf9PBnKJUpNWctEFAp5q9LObX353tIkArBdsA,639
+entorno/bin/rst2html4.py,sha256=z0108x1oxTOXg-oU_s4j9cv7OxCrPTTmV6LfADMtbjs,761
+entorno/bin/rst2html5.py,sha256=qMrEoZ1mjo5Tt8yCqXNCquYJ-DDynFVqKN5ZeOfLLB8,1096
+entorno/bin/rst2latex.py,sha256=8gZ9ryotGng3K0goX5P51-L3oKW9chJXpQumAPJkwyU,838
+entorno/bin/rst2man.py,sha256=ryMvgZX7IDfC7N-vTywTj-8iVlpHJkbMTEyDhHIdOsI,661
+entorno/bin/rst2odt.py,sha256=Mjs19kOZ30tHB1cRb9km3anspnLv7ftoSlw2zHdpvvw,827
+entorno/bin/rst2odt_prepstyles.py,sha256=8G8wdgn7UhaxI-PZYLVIwssvtSUSt1vmEuOI1HCZlrc,633
+entorno/bin/rst2pseudoxml.py,sha256=ydQMcz9e0Zu6O8go0I_QoG7t8CtbRAzXy5I27JJcIV4,646
+entorno/bin/rst2s5.py,sha256=4uDPt_H4qIMkQVLcXxVgIyy01Evhv34BBLEMVFzZM48,682
+entorno/bin/rst2xetex.py,sha256=ISxJ2axfgO94mR6a7n6eFC1jMBiwwqZUFLROX5BjVJs,918
+entorno/bin/rst2xml.py,sha256=vfsZFTKMpkJ1TZuttq2UrKgFWDOwOd8mWG8qHbu_6QA,647
+entorno/bin/rstpep2html.py,sha256=jrbcAcQDtHW6S1f9Ef4XM0DuzbDG-VuE2I_AeRAelGs,715
+experiments/scripts/compute_metrics.py,sha256=vBfOg2E967AhRqZQhoASNcDcZTVTjgobyW2U4UAfHt0,8047
+experiments/scripts/metrics_complexity_analysis.py,sha256=ky5tNiiqhGazDCJr_QiFegD4inEiIM6rCwgMFBXxKNc,3921
+experiments/scripts/metro_experiment.py,sha256=3xXE1FWj88HnJ2vWL14n-1GoEKZX218NG_PNEXBkAOs,5051
+experiments/scripts/opt_metro_experiment.py,sha256=NAfmzmPPh_malkmtTRstByJoRbJnrC8UyWQFuRJ6-1g,13011
+tests/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+tests/test_binary.py,sha256=3uLdXzdbQcqLRfXeKDmt2g2XQTU5lZOFQoiy-r9Olqo,29801
+tests/test_non_binary.py,sha256=PRvsgGbwbRXVNkYBk2-oJw8ekLW4DWJkEZhAWO0L9sk,11309
+tsadmetrics/__init__.py,sha256=MTWOa43fgOdkMNo5NglCReRnB8hoF0ob2PIvDziCNHw,1575
+tsadmetrics/binary_metrics.py,sha256=nwfPdfHAc_4tJMNlyIwMwFQRLvCU-ik9lQLqlaWLqTs,37741
+tsadmetrics/metric_utils.py,sha256=Y_lOE01_uyC22wnw3_G-kKUEJdqevDIWMWvSDE8Cjms,10477
+tsadmetrics/non_binary_metrics.py,sha256=JIOvkigSjHBZLKbGJj7ESe0lPM7P1JPoIUnbiMZuuLg,2896
+tsadmetrics/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+tsadmetrics/utils.py,sha256=G0yWxgTZ9MBzyB0XDLrO2TMwmtm4hssDp5Sr0CG9FqY,1834
+tsadmetrics/_tsadeval/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+tsadmetrics/_tsadeval/auc_roc_pr_plot.py,sha256=PHqJUXq2qI248XV9o04D8SsUJgowetaKq0Cu5bYrIAE,12689
+tsadmetrics/_tsadeval/discontinuity_graph.py,sha256=Ci65l_DPi6HTtb8NvQJe1najgGrRuEpOMWvSyi2AeR0,4088
+tsadmetrics/_tsadeval/latency_sparsity_aware.py,sha256=92wt6ARSXL5Y-281joNaSu1E7hnkIbl3m6gyzODTYBE,12092
+tsadmetrics/_tsadeval/metrics.py,sha256=d-1VpJu_mp8gZjW2FeD7eqkFKEkGsYcsy6DcSGK4kSk,24100
+tsadmetrics/_tsadeval/nabscore.py,sha256=8H4cgzzjXrbQzpI-YKEJj31eSGSROrT7NNC86n9d2yY,11696
+tsadmetrics/_tsadeval/tests.py,sha256=KjFPlEHWYkxHXtaEs1_WiTgATEtr7kPKQbgsljSxJ8o,12697
+tsadmetrics/_tsadeval/threshold_plt.py,sha256=ExgxIcsDMmgLNveNug5fimEhEe6Km0g68npQj-7oWOE,726
+tsadmetrics/_tsadeval/time_tolerant.py,sha256=duq3B58ohjS6QkWdNUuCQFt2xmCJ0dMWTVzOr6E3H0A,1486
+tsadmetrics/_tsadeval/vus_utils.py,sha256=XL5tV9hxBW8aGkobT84cp2FdHNuNZ3PUgaplwHsDjNk,7868
+tsadmetrics/_tsadeval/affiliation/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+tsadmetrics/_tsadeval/affiliation/_affiliation_zone.py,sha256=Oyib8Rzr5hSvq-NTyHPekLqd6_iDGu-Pc1MocNzgG6I,3486
+tsadmetrics/_tsadeval/affiliation/_integral_interval.py,sha256=gXfmnsywrk6nay9U9nma5hdYtaEDYD-uPRo6jTMefNo,20845
+tsadmetrics/_tsadeval/affiliation/_single_ground_truth_event.py,sha256=oEtdqg-w72iJOAfHq11DU7Q_DVIxrgbvAAay5s-TkWU,3488
+tsadmetrics/_tsadeval/affiliation/generics.py,sha256=FnQ7IE0k_ri6neqrzrewz5AFIBNuijt35JlEj8qvv8k,4882
+tsadmetrics/_tsadeval/affiliation/metrics.py,sha256=EZLnHI2ACpPzB8HuqTRkFXHJuKa1jJ0jGa1BZyDqC0Q,5045
+tsadmetrics/_tsadeval/eTaPR_pkg/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+tsadmetrics/_tsadeval/eTaPR_pkg/etapr.py,sha256=3CJRdDL-edxKrHGmQHPzZDg8X5O42NGLXme6nc-jfgE,15768
+tsadmetrics/_tsadeval/eTaPR_pkg/tapr.py,sha256=KGgRyLVBcHQfDyaPtyU8t3Iq1CZreIkc6xJ8BZfygQw,14645
+tsadmetrics/_tsadeval/eTaPR_pkg/DataManage/File_IO.py,sha256=98Nkr84vvZ0rpLIcH28DrR7RaBistLpF_8rZ9MpUj-0,6691
+tsadmetrics/_tsadeval/eTaPR_pkg/DataManage/Range.py,sha256=ZJuaX3kNtCuZNmOAxIXJejawoeZrH6hupX3DILIqmCM,1825
+tsadmetrics/_tsadeval/eTaPR_pkg/DataManage/Time_Plot.py,sha256=IIrMdeM4YUrNmUskLw58crCN8URqY8rHn7q5zJDjDUU,8497
+tsadmetrics/_tsadeval/eTaPR_pkg/DataManage/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+tsadmetrics/_tsadeval/prts/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+tsadmetrics/_tsadeval/prts/basic_metrics_ts.py,sha256=GPM2EdLL4tmxuuD6Q15Uk9IuAFPkSMn8LZpxLTYao_0,6066
+tsadmetrics/_tsadeval/prts/base/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+tsadmetrics/_tsadeval/prts/base/time_series_metrics.py,sha256=ngKaP_8IwmA6GzDlGYSE32yDNoj7wmN9rgB-EeKwvG0,5615
+tsadmetrics/_tsadeval/prts/time_series_metrics/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+tsadmetrics/_tsadeval/prts/time_series_metrics/fscore.py,sha256=pJz4iuPyVGNvwsaRY9F0DlCbKAKJv4CdZl8Ulvyq6pw,2336
+tsadmetrics/_tsadeval/prts/time_series_metrics/precision.py,sha256=jLkcMg7UNl25SHtZUBGkP-RV8HsvaZCtjakryl7PFWU,3204
+tsadmetrics/_tsadeval/prts/time_series_metrics/precision_recall.py,sha256=OhUJSm_I7VZ_gX_SSg8AYUq3_NW9rMIy7lAVsnOFw4Q,417
+tsadmetrics/_tsadeval/prts/time_series_metrics/recall.py,sha256=LL-0pPer3ymovVRlktaHo5XDzpgiDhWOVfdPOzKR6og,3152
+tsadmetrics-0.1.5.dist-info/METADATA,sha256=Z1lOZK8SyydVHNA_-JKWuoogGAw-apiynWXIkc9vGvk,756
+tsadmetrics-0.1.5.dist-info/WHEEL,sha256=iAkIy5fosb7FzIOwONchHf19Qu7_1wCWyFNR5gu9nU0,91
+tsadmetrics-0.1.5.dist-info/top_level.txt,sha256=RBlddXJB5cEhcrINYC_6YGVtpv4GXGan4eVNyobrVRk,38
+tsadmetrics-0.1.5.dist-info/RECORD,,

tsadmetrics-0.1.5.dist-info/top_level.txt

@@ -0,0 +1,4 @@
+entorno
+experiments
+tests
+tsadmetrics

tsadmetrics-0.1.4.dist-info/RECORD

@@ -1,20 +0,0 @@
-tsadmetrics/__init__.py,sha256=MTWOa43fgOdkMNo5NglCReRnB8hoF0ob2PIvDziCNHw,1575
-tsadmetrics/binary_metrics.py,sha256=nwfPdfHAc_4tJMNlyIwMwFQRLvCU-ik9lQLqlaWLqTs,37741
-tsadmetrics/metric_utils.py,sha256=Y_lOE01_uyC22wnw3_G-kKUEJdqevDIWMWvSDE8Cjms,10477
-tsadmetrics/non_binary_metrics.py,sha256=JIOvkigSjHBZLKbGJj7ESe0lPM7P1JPoIUnbiMZuuLg,2896
-tsadmetrics/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-tsadmetrics/utils.py,sha256=G0yWxgTZ9MBzyB0XDLrO2TMwmtm4hssDp5Sr0CG9FqY,1834
-tsadmetrics/_tsadeval/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-tsadmetrics/_tsadeval/auc_roc_pr_plot.py,sha256=PHqJUXq2qI248XV9o04D8SsUJgowetaKq0Cu5bYrIAE,12689
-tsadmetrics/_tsadeval/discontinuity_graph.py,sha256=Ci65l_DPi6HTtb8NvQJe1najgGrRuEpOMWvSyi2AeR0,4088
-tsadmetrics/_tsadeval/latency_sparsity_aware.py,sha256=92wt6ARSXL5Y-281joNaSu1E7hnkIbl3m6gyzODTYBE,12092
-tsadmetrics/_tsadeval/metrics.py,sha256=d-1VpJu_mp8gZjW2FeD7eqkFKEkGsYcsy6DcSGK4kSk,24100
-tsadmetrics/_tsadeval/nabscore.py,sha256=8H4cgzzjXrbQzpI-YKEJj31eSGSROrT7NNC86n9d2yY,11696
-tsadmetrics/_tsadeval/tests.py,sha256=KjFPlEHWYkxHXtaEs1_WiTgATEtr7kPKQbgsljSxJ8o,12697
-tsadmetrics/_tsadeval/threshold_plt.py,sha256=ExgxIcsDMmgLNveNug5fimEhEe6Km0g68npQj-7oWOE,726
-tsadmetrics/_tsadeval/time_tolerant.py,sha256=duq3B58ohjS6QkWdNUuCQFt2xmCJ0dMWTVzOr6E3H0A,1486
-tsadmetrics/_tsadeval/vus_utils.py,sha256=XL5tV9hxBW8aGkobT84cp2FdHNuNZ3PUgaplwHsDjNk,7868
-tsadmetrics-0.1.4.dist-info/METADATA,sha256=7CclFJyIyGpieaaykLuCaTh-P1u4O0KnOrdgL8ZQoWk,756
-tsadmetrics-0.1.4.dist-info/WHEEL,sha256=iAkIy5fosb7FzIOwONchHf19Qu7_1wCWyFNR5gu9nU0,91
-tsadmetrics-0.1.4.dist-info/top_level.txt,sha256=rRMFvkwJRUuenl0__YY_3BNr-rkdhAdj28iICkpC5a4,12
-tsadmetrics-0.1.4.dist-info/RECORD,,

tsadmetrics-0.1.4.dist-info/top_level.txt

@@ -1 +0,0 @@
-tsadmetrics