Rename the two different ChangePoint classes for clarity

otava/analysis.py

@@ -15,6 +15,7 @@
 # specific language governing permissions and limitations
 # under the License.
 
+import copy
 from dataclasses import dataclass
 from typing import List, Optional, Sequence, SupportsFloat, Tuple
 
@@ -24,7 +25,7 @@
 from otava.change_point_divisive.base import (
     BaseStats,
     CandidateChangePoint,
-    ChangePoint,
+    ChangePointOtava,
     GenericStats,
     SignificanceTester,
 )
@@ -97,11 +98,11 @@ def to_json(self):
 
 
 # Generic Change Point List
-GenCPList = List[ChangePoint[GenericStats]]
+GenCPList = List[ChangePointOtava[GenericStats]]
 # Permutation Change Point List
-PermCPList = List[ChangePoint[PermutationStats]]
+PermCPList = List[ChangePointOtava[PermutationStats]]
 # T-test Change Point List
-TtestCPList = List[ChangePoint[TTestStats]]
+TtestCPList = List[ChangePointOtava[TTestStats]]
 
 
 class TTestSignificanceTester(SignificanceTester):
@@ -130,7 +131,7 @@ def compare(self, left: Sequence[SupportsFloat], right: Sequence[SupportsFloat])
 
     def change_point(
         self, candidate: CandidateChangePoint, series: Sequence[SupportsFloat], intervals: List[slice]
-    ) -> ChangePoint[TTestStats]:
+    ) -> ChangePointOtava[TTestStats]:
         """
         Computes properties of the change point if the Candidate Change Point based on the provided intervals.
 
@@ -167,7 +168,7 @@ def change_point(
         left = series[left_interval]
         right = series[right_interval]
         stats = self.compare(left, right)
-        return ChangePoint.from_candidate(candidate, stats)
+        return ChangePointOtava.from_candidate(candidate, stats)
 
 
 def fill_missing(data: Sequence[SupportsFloat]):
@@ -323,4 +324,4 @@ def compute_change_points(
     """
     first_pass_pvalue = max_pvalue * 10 if max_pvalue < 0.05 else (max_pvalue * 2 if max_pvalue < 0.5 else max_pvalue)
     weak_change_points = split(series, window_len, first_pass_pvalue, new_points=new_data, old_cp=old_weak_cp)
-    return merge(weak_change_points, series, max_pvalue, min_magnitude), weak_change_points
+    return merge(copy.copy(weak_change_points), series, max_pvalue, min_magnitude), weak_change_points

otava/bigquery.py

@@ -22,7 +22,7 @@
 from google.cloud import bigquery
 from google.oauth2 import service_account
 
-from otava.analysis import ChangePoint
+from otava.analysis import ChangePointOtava
 from otava.test_config import BigQueryTestConfig
 
 
@@ -87,7 +87,7 @@ def insert_change_point(
         test: BigQueryTestConfig,
         metric_name: str,
         attributes: Dict,
-        change_point: ChangePoint,
+        change_point: ChangePointOtava,
     ):
         kwargs = {**attributes, **{test.time_column: datetime.utcfromtimestamp(change_point.time)}}
         update_stmt = test.update_stmt.format(

otava/change_point_divisive/base.py

@@ -39,7 +39,7 @@ class BaseStats:
 
 
 @dataclass
-class ChangePoint(CandidateChangePoint, Generic[GenericStats]):
+class ChangePointOtava(CandidateChangePoint, Generic[GenericStats]):
     '''Change point class, defined by index and signigicance test statistic.'''
     stats: GenericStats
 
@@ -48,7 +48,7 @@ def __eq__(self, other):
         return isinstance(other, self.__class__) and self.index == other.index
 
     @classmethod
-    def from_candidate(cls, candidate: CandidateChangePoint, stats: GenericStats) -> 'ChangePoint[GenericStats]':
+    def from_candidate(cls, candidate: CandidateChangePoint, stats: GenericStats) -> 'ChangePointOtava[GenericStats]':
         return cls(
             index=candidate.index,
             qhat=candidate.qhat,
@@ -67,7 +67,7 @@ class SignificanceTester(Generic[GenericStats]):
     def __init__(self, max_pvalue: float):
         self.max_pvalue = max_pvalue
 
-    def get_intervals(self, change_points: List[ChangePoint[GenericStats]]) -> List[slice]:
+    def get_intervals(self, change_points: List[ChangePointOtava[GenericStats]]) -> List[slice]:
         '''Returns list of slices of the series. Change points must be sorted by index.'''
         assert all(
             change_points[i].index <= change_points[i + 1].index
@@ -82,12 +82,12 @@ def get_intervals(self, change_points: List[ChangePoint[GenericStats]]) -> List[
         ]
         return [interval for interval in intervals if interval.start != interval.stop]
 
-    def is_significant(self, point: ChangePoint[GenericStats]) -> bool:
-        '''Compares ChangePoint to level of significance max_pvalue'''
+    def is_significant(self, point: ChangePointOtava[GenericStats]) -> bool:
+        '''Compares ChangePointOtava to level of significance max_pvalue'''
         return point.stats.pvalue <= self.max_pvalue
 
-    def change_point(self, candidate: CandidateChangePoint, series: NDArray, intervals: List[slice]) -> ChangePoint[GenericStats]:
-        '''Computes stats for a change point candidate and wraps it into ChangePoint class'''
+    def change_point(self, candidate: CandidateChangePoint, series: NDArray, intervals: List[slice]) -> ChangePointOtava[GenericStats]:
+        '''Computes stats for a change point candidate and wraps it into ChangePointOtava class'''
         ...
 
 

otava/change_point_divisive/detector.py

@@ -21,7 +21,7 @@
 
 from otava.change_point_divisive.base import (
     Calculator,
-    ChangePoint,
+    ChangePointOtava,
     GenericStats,
     SignificanceTester,
 )
@@ -32,7 +32,7 @@ def __init__(self, significance_tester: SignificanceTester, calculator: Type[Cal
         self.tester = significance_tester
         self.calculator = calculator
 
-    def get_change_points(self, series: Sequence[SupportsFloat], start: Optional[int] = None, end: Optional[int] = None) -> List[ChangePoint[GenericStats]]:
+    def get_change_points(self, series: Sequence[SupportsFloat], start: Optional[int] = None, end: Optional[int] = None) -> List[ChangePointOtava[GenericStats]]:
         '''Finds change points in `series[start : end]`.'''
         if not isinstance(series, np.ndarray):
             series = np.array(series[start : end], dtype=np.float64)

otava/change_point_divisive/significance_test.py

@@ -25,7 +25,7 @@
     BaseStats,
     Calculator,
     CandidateChangePoint,
-    ChangePoint,
+    ChangePointOtava,
     SignificanceTester,
 )
 
@@ -50,7 +50,7 @@ def __init__(self, max_pvalue: float, permutations: int, calculator: Type[Calcul
         self.seed = seed
         self.rng = np.random.default_rng(seed)
 
-    def change_point(self, candidate: CandidateChangePoint, series: NDArray, intervals: List[slice]) -> ChangePoint[PermutationStats]:
+    def change_point(self, candidate: CandidateChangePoint, series: NDArray, intervals: List[slice]) -> ChangePointOtava[PermutationStats]:
         '''Perform permutation test within candidate cluster'''
 
         # 1. Find permutated Qhats
@@ -74,4 +74,4 @@ def change_point(self, candidate: CandidateChangePoint, series: NDArray, interva
             extreme_qhat_perm=extreme_qhat_perm,
             n_perm=self.permutations
         )
-        return ChangePoint.from_candidate(candidate, stats)
+        return ChangePointOtava.from_candidate(candidate, stats)

otava/postgres.py

@@ -21,7 +21,7 @@
 
 import pg8000
 
-from otava.analysis import ChangePoint
+from otava.analysis import ChangePointOtava
 from otava.test_config import PostgresTestConfig
 
 
@@ -88,7 +88,7 @@ def insert_change_point(
         test: PostgresTestConfig,
         metric_name: str,
         attributes: Dict,
-        change_point: ChangePoint,
+        change_point: ChangePointOtava,
     ):
         cursor = self.__get_conn().cursor()
         kwargs = {**attributes, **{test.time_column: datetime.utcfromtimestamp(change_point.time)}}

otava/series.py

@@ -27,7 +27,7 @@
     compute_change_points_orig,
     fill_missing,
 )
-from otava.change_point_divisive.base import ChangePoint as _ChangePoint
+from otava.change_point_divisive.base import ChangePointOtava
 
 
 @dataclass
@@ -72,7 +72,7 @@ def to_json(self):
 
 
 @dataclass
-class ChangePoint(_ChangePoint[TTestStats]):
+class ChangePointHunter(ChangePointOtava[TTestStats]):
     """A change-point for a single metric"""
     metric: str
     time: int
@@ -140,7 +140,7 @@ class ChangePointGroup:
     prev_time: int
     attributes: Dict[str, str]
     prev_attributes: Dict[str, str]
-    changes: List[ChangePoint]
+    changes: List[ChangePointHunter]
 
     def to_json(self, rounded=False):
         return {
@@ -215,11 +215,11 @@ class AnalyzedSeries:
 
     __series: Series
     options: AnalysisOptions
-    change_points: Dict[str, List[ChangePoint]]
+    change_points: Dict[str, List[ChangePointHunter]]
     change_points_by_time: List[ChangePointGroup]
     change_points_timestamp: Any
 
-    def __init__(self, series: Series, options: AnalysisOptions, change_points: Dict[str, ChangePoint] = None):
+    def __init__(self, series: Series, options: AnalysisOptions, change_points: Dict[str, ChangePointHunter] = None):
         self.__series = series
         self.options = options
         self.change_points_timestamp = datetime.now(tz=timezone.utc)
@@ -235,7 +235,7 @@ def __init__(self, series: Series, options: AnalysisOptions, change_points: Dict
     @staticmethod
     def __compute_change_points(
         series: Series, options: AnalysisOptions
-    ) -> Dict[str, List[ChangePoint]]:
+    ) -> Dict[str, List[ChangePointHunter]]:
         result = {}
         weak_change_points = {}
         for metric in series.data.keys():
@@ -258,13 +258,13 @@ def __compute_change_points(
                 )
                 for c in weak_cps:
                     weak_change_points[metric].append(
-                        ChangePoint(
+                        ChangePointHunter(
                             index=c.index, qhat=0.0, time=series.time[c.index], metric=metric, stats=c.stats
                         )
                     )
                 for c in change_points:
                     result[metric].append(
-                        ChangePoint(
+                        ChangePointHunter(
                             index=c.index, qhat=0.0, time=series.time[c.index], metric=metric, stats=c.stats
                         )
                     )
@@ -274,9 +274,9 @@ def __compute_change_points(
 
     @staticmethod
     def __group_change_points_by_time(
-        series: Series, change_points: Dict[str, List[ChangePoint]]
+        series: Series, change_points: Dict[str, List[ChangePointHunter]]
     ) -> List[ChangePointGroup]:
-        changes: List[ChangePoint] = []
+        changes: List[ChangePointHunter] = []
         for metric in change_points.keys():
             changes += change_points[metric]
 
@@ -381,14 +381,14 @@ def append(self, time, new_data, attributes):
             result[metric] = []
             for c in change_points:
                 result[metric].append(
-                    ChangePoint(
+                    ChangePointHunter(
                         index=c.index, qhat=0.0, time=self.__series.time[c.index], metric=metric, stats=c.stats
                     )
                 )
             weak_change_points[metric] = []
             for c in weak_cps:
                 weak_change_points[metric].append(
-                    ChangePoint(
+                    ChangePointHunter(
                         index=c.index, qhat=0.0, time=self.__series.time[c.index], metric=metric, stats=c.stats
                     )
                 )
@@ -493,7 +493,7 @@ def from_json(cls, analyzed_json):
                     pvalue=cp["pvalue"],
                 )
                 new_list.append(
-                    ChangePoint(
+                    ChangePointHunter(
                         index=cp["index"], time=cp["time"], metric=cp["metric"], stats=stat
                     )
                 )
@@ -511,7 +511,7 @@ def from_json(cls, analyzed_json):
                     pvalue=cp["pvalue"],
                 )
                 new_list.append(
-                    ChangePoint(
+                    ChangePointHunter(
                         index=cp["index"], time=cp["time"], metric=cp["metric"], stats=stat
                     )
                 )

tests/change_point_divisive_test.py

@@ -19,7 +19,7 @@
 import pytest
 
 from otava.analysis import TTestSignificanceTester, TTestStats
-from otava.change_point_divisive.base import ChangePoint
+from otava.change_point_divisive.base import ChangePointOtava
 from otava.change_point_divisive.calculator import PairDistanceCalculator
 from otava.change_point_divisive.detector import ChangePointDetector
 from otava.change_point_divisive.significance_test import PermutationsSignificanceTester
@@ -124,9 +124,9 @@ def test_get_intervals_requires_sorted_change_points():
 
     # Sorted change points should work
     sorted_cps = [
-        ChangePoint(index=5, qhat=1.0, stats=stats),
-        ChangePoint(index=10, qhat=1.0, stats=stats),
-        ChangePoint(index=15, qhat=1.0, stats=stats),
+        ChangePointOtava(index=5, qhat=1.0, stats=stats),
+        ChangePointOtava(index=10, qhat=1.0, stats=stats),
+        ChangePointOtava(index=15, qhat=1.0, stats=stats),
     ]
     intervals = tester.get_intervals(sorted_cps)
     assert len(intervals) == 4
@@ -137,9 +137,9 @@ def test_get_intervals_requires_sorted_change_points():
 
     # Unsorted change points should raise AssertionError
     unsorted_cps = [
-        ChangePoint(index=10, qhat=1.0, stats=stats),
-        ChangePoint(index=5, qhat=1.0, stats=stats),
-        ChangePoint(index=15, qhat=1.0, stats=stats),
+        ChangePointOtava(index=10, qhat=1.0, stats=stats),
+        ChangePointOtava(index=5, qhat=1.0, stats=stats),
+        ChangePointOtava(index=15, qhat=1.0, stats=stats),
     ]
     with pytest.raises(AssertionError, match="Change points must be sorted by index"):
         tester.get_intervals(unsorted_cps)