Merge pull request #831 from GabrielKS/fix_confidence_threshold

Get confidenceThreshold from config file

emission/analysis/configs/expectation_notification_config.py

@@ -77,7 +77,7 @@ def _mode_matches_date(mode, trip_date):
 
 def _get_collection_mode_by_schedule(trip):
     trip_tz = trip["data"]["end_local_dt"]["timezone"]
-    trip_date = arrow.get(trip["data"]["end_ts"], tz=trip_tz)
+    trip_date = arrow.get(trip["data"]["end_ts"]).to(trip_tz)
     for mode in _config["modes"]:
         if _mode_matches_date(mode, trip_date): return mode
     raise ValueError("Trip date does not match any modes; this means the config file lacks a schedule-less mode")

emission/analysis/userinput/expectations.py

@@ -38,10 +38,12 @@ def _process_and_save_trip(user_id, inferred_trip, ts):
     if _test_options["preprocess_trip"] is not None: _test_options["preprocess_trip"](expected_trip)
 
     expectation = _get_expectation_for_trip(expected_trip)
+    confidence_threshold = eace.get_confidence_threshold(expected_trip)
     # For now, I don't think it's necessary to save each expectation as its own database entry
 
     expected_trip["data"]["inferred_trip"] = inferred_trip.get_id()
     expected_trip["data"]["expectation"] = expectation
+    expected_trip["data"]["confidence_threshold"] = confidence_threshold
     ts.insert(expected_trip)
 
 # This is a placeholder. TODO: implement the real algorithm

emission/core/wrapper/confirmedtrip.py

@@ -15,6 +15,7 @@ class Confirmedtrip(ecwt.Trip):
                   "inferred_labels": ecwb.WrapperBase.Access.WORM,
                   "inferred_trip": ecwb.WrapperBase.Access.WORM,
                   "expectation": ecwb.WrapperBase.Access.WORM,
+                  "confidence_threshold": ecwb.WrapperBase.Access.WORM,
                   "expected_trip": ecwb.WrapperBase.Access.WORM,
 # the confirmed section that is the "primary"
 # https://github.com/e-mission/e-mission-docs/issues/476#issuecomment-738120752

emission/core/wrapper/expectedtrip.py

@@ -14,7 +14,8 @@ class Expectedtrip(ecwt.Trip):
                   "cleaned_trip": ecwb.WrapperBase.Access.WORM,
                   "inferred_labels": ecwb.WrapperBase.Access.WORM,
                   "inferred_trip": ecwb.WrapperBase.Access.WORM,
-                  "expectation": ecwb.WrapperBase.Access.WORM
+                  "expectation": ecwb.WrapperBase.Access.WORM,
+                  "confidence_threshold": ecwb.WrapperBase.Access.WORM,
                   })
 
     def _populateDependencies(self):

emission/tests/pipelineTests/TestExpectationNotificationConfig.py

@@ -16,7 +16,7 @@ def setUp(self):
         }
         eace.reload_config()
 
-        self.tz = "America/Chicago"
+        self.tz = "Etc/GMT-8"
         # Note that these depend on certain values in expectations.conf.json.sample, as will other values later
         self.test_dates = {
             "before_intensive": arrow.get("2021-05-01T20:00:00.000", tzinfo=self.tz),  # Ensures schedules can't apply before their start_date

emission/tests/pipelineTests/TestExpectationPipeline.py

@@ -26,7 +26,19 @@ class TestExpectationPipeline(unittest.TestCase):
         960: arrow.get("2023-02-12T20:00:00.000", tzinfo=tz) 
     }
 
+    @staticmethod
+    def fingerprint(trip):
+        # See eacilp.placeholder_predictor_2 for an explanation of the "fingerprint" technique
+        return trip["data"]["start_local_dt"]["hour"]*60+trip["data"]["start_local_dt"]["minute"]
+
     def setUp(self):
+        self.test_options_stash = eace._test_options
+        eace._test_options = {
+            "use_sample": True,
+            "override_keylist": None
+        }
+        eace.reload_config()
+
         np.random.seed(61297777)
         self.reset_all()
         etc.setupRealExample(self, "emission/tests/data/real_examples/shankari_2015-07-22")
@@ -38,6 +50,9 @@ def setUp(self):
 
     def tearDown(self):
         self.reset_all()
+
+        eace._test_options = self.test_options_stash
+        eace.reload_config()
 
     def run_pipeline(self, algorithms):
         primary_algorithms_stash = eacilp.primary_algorithms
@@ -49,9 +64,7 @@ def run_pipeline(self, algorithms):
         eaue._test_options = test_options_stash
 
     def preprocess(self, trip):
-        # See eacilp.placeholder_predictor_2 for an explanation of the "fingerprint" technique
-        fingerprint = trip["data"]["start_local_dt"]["hour"]*60+trip["data"]["start_local_dt"]["minute"]
-        trip["data"]["end_ts"] = self.contrived_dates[fingerprint].float_timestamp
+        trip["data"]["end_ts"] = self.contrived_dates[self.fingerprint(trip)].float_timestamp
         trip["data"]["end_local_dt"]["timezone"] = self.tz
 
     def reset_all(self):
@@ -79,8 +92,7 @@ def testRawAgainstAnswers(self):
             960: {"type": "randomFraction", "value": 0.05}
         }
         for trip in self.expected_trips:
-            fingerprint = trip["data"]["start_local_dt"]["hour"]*60+trip["data"]["start_local_dt"]["minute"]
-            self.assertEqual(eace.get_expectation(trip), answers[fingerprint])
+            self.assertEqual(eace.get_expectation(trip), answers[self.fingerprint(trip)])
 
     def testProcessedAgainstAnswers(self):
         answers = {
@@ -92,21 +104,34 @@ def testProcessedAgainstAnswers(self):
             960: None
         }
         for trip in self.expected_trips:
-            fingerprint = trip["data"]["start_local_dt"]["hour"]*60+trip["data"]["start_local_dt"]["minute"]
-            if answers[fingerprint] is not None: self.assertEqual(trip["data"]["expectation"]["to_label"], answers[fingerprint])
+            ans = answers[self.fingerprint(trip)]
+            if ans is not None: self.assertEqual(trip["data"]["expectation"]["to_label"], ans)
 
     def testProcessedAgainstRaw(self):
         for trip in self.expected_trips:
             self.assertIn("expectation", trip["data"])
             raw_expectation = eace.get_expectation(trip)
             if raw_expectation["type"] == "none":
-                self.assertEqual(trip["data"]["expectation"], {"to_label": False})
+                self.assertEqual(trip["data"]["expectation"]["to_label"], False)
             elif raw_expectation["type"] == "all":
-                self.assertEqual(trip["data"]["expectation"], {"to_label": True})
+                self.assertEqual(trip["data"]["expectation"]["to_label"], True)
             else:
                 print("Expectation behavior for "+str(raw_expectation)+" has not been implemented yet; not testing. Value is "+str(trip["data"]["expectation"]))
             # TODO: implement tests for the other configurable expectation types once they've been implemented
 
+    def testConfidenceThreshold(self):
+        answers = {
+            494: 0.55,
+            565: 0.65,
+            795: 0.55,
+            805: 0.65,
+            880: 0.65,
+            960: 0.55
+        }
+        for trip in self.expected_trips:
+            self.assertTrue("confidence_threshold" in trip["data"])  # Existence
+            self.assertAlmostEqual(trip["data"]["confidence_threshold"], answers[self.fingerprint(trip)]) # Correctness
+
 def main():
     etc.configLogging()
     unittest.main()