adjust date parsing for tz aware and naive inputs

app/src/data_loading/main.py

@@ -154,8 +154,7 @@ def _get_active_time_limit(self) -> datetime:
         """
         Get the current threshold for marking sensor as active.
         """
-        limit = pd.Timestamp("now") + pd.Timedelta(-4, unit="H")
-        limit += pd.Timedelta(-1, unit="H")
+        limit = pd.Timestamp("now", tz='EST') + pd.Timedelta(-2, unit="H")
 
         return limit
 

app/src/utils.py

@@ -234,8 +234,7 @@ def _enum_col_names_to_string(df: pd.DataFrame) -> pd.DataFrame:
 
         return new_df
 
-    @staticmethod
-    def _set_data_types(df: pd.DataFrame) -> pd.DataFrame:
+    def _set_data_types(self, df: pd.DataFrame) -> pd.DataFrame:
         """
         Sets the right data types for noise data columns in place.
         """
@@ -251,13 +250,27 @@ def _set_data_types(df: pd.DataFrame) -> pd.DataFrame:
             if col in df.columns:
                 df[col] = df[col].astype(type_)
 
-        date_cols = [COLUMN.TIMESTAMP, COLUMN.LATEST_TIMESTAMP]
-        for col in date_cols:
+        tz_aware_date_cols = [COLUMN.TIMESTAMP, COLUMN.LATEST_TIMESTAMP]
+        for col in tz_aware_date_cols:
+            if col in df.columns:
+                df[col] = self._convert_tz_aware_to_est(df[col])
+
+        tz_naive_date_cols = [COLUMN.START, COLUMN.END]
+        for col in tz_naive_date_cols:
             if col in df.columns:
-                df[col] = pd.to_datetime(df[col]).dt.tz_localize(None)
+                print(df[col])
+                df[col] = self._convert_tz_naive_to_est(df[col])
 
         return df
 
+    @staticmethod
+    def _convert_tz_naive_to_est(datetime_col: pd.Series) -> pd.Series:
+        return pd.to_datetime(datetime_col).dt.tz_localize('EST')
+
+    @staticmethod
+    def _convert_tz_aware_to_est(datetime_col: pd.Series) -> pd.Series:
+        return pd.to_datetime(datetime_col).dt.tz_convert('EST')
+
     @staticmethod
     def _raw_to_dataframe(raw_data: List[Dict[str, Any]]) -> pd.DataFrame:
         """

app/tests/test_data_processing.py

@@ -1,8 +1,20 @@
 import pandas as pd
 from src.utils import DataFormatter, COLUMN, date_to_string
+from src.data_loading.main import AppDataManager
 import pytest
 from datetime import datetime, date
 
+def test_time_comparison():
+    data_manager = AppDataManager()
+    limit = data_manager._get_active_time_limit()
+
+    column = pd.Series(["2024-05-18T11:00:00-04:00"])
+
+    data_formatter = DataFormatter()
+    column = data_formatter._convert_tz_aware_to_est(column)
+
+    assert (column < limit).all()
+
 
 def date_to_string():
     """

makefile

@@ -132,4 +132,8 @@ debug:
 #############
 
 count:
-	wc -l **/*.py makefile Dockerfile README.md
+	wc -l **/*.py makefile Dockerfile README.md
+
+remove_branch:
+	git push -d origin $(BRANCH)
+	git branch -d $(BRANCH)