Truncate real world data to last sampling time

pseudobatch/datasets/_dataloaders.py

@@ -91,7 +91,20 @@ def load_real_world_yeast_fedbatch():
     is obtained from an experiment carried out in a biolector."""
 
     data_path = pathlib.Path(__file__).parent / "data" / "biolector_yeast_fedbatch.csv"
-    return pd.read_csv(data_path)
+    data = pd.read_csv(data_path)
+
+    def truncate_after_last_sample(
+        df: pd.DataFrame, time_col: str, sample_col: str
+    ) -> pd.DataFrame:
+        last_sample_time = df.query(f"`{sample_col}` > 0")[time_col].max()
+
+        return df.query(f"`{time_col}` <= @last_sample_time")
+
+    data_truncated = data.groupby("Biolector well").apply(
+        lambda x: truncate_after_last_sample(x, "Time", "Sample volume")
+    )
+
+    return data_truncated.reset_index(drop=True)
 
 
 def load_volatile_compounds_fedbatch(sampling_points_only: bool = False):
@@ -107,4 +120,4 @@ def load_volatile_compounds_fedbatch(sampling_points_only: bool = False):
         If True, only the rows where a sample was taken is kept, by default False
     """
     data_path = pathlib.Path(__file__).parent / "data" / "volatile_product.csv"
-    return _prepare_simulated_dataset(data_path, sampling_points_only=sampling_points_only)
+    return _prepare_simulated_dataset(data_path, sampling_points_only=sampling_points_only)

tests/test_unit.py

@@ -42,7 +42,7 @@ def test_load_real_world_yeast_fedbatch():
     df = load_real_world_yeast_fedbatch()
     logging.debug(df.shape)
     assert df.empty is False
-    assert df.shape == (11400, 12), "The dataset has changed. Update the test."
+    assert df.shape == (10712, 12), "The dataset has changed. Update the test."
 
 
 def test_load_volatile_compounds_fedbatch_unique_timestamps():