fixup! Support parquet observation files in SeismicObservation

Author: akustn

src/ert/config/_observations.py

@@ -1076,9 +1076,9 @@ def from_obs_dict(
     ) -> list[Self]:
         """Create seismic observations from an observation dictionary.
 
-        File containing seismic observations (CSV) and file containing the polygon
-        limiting observations used in the match step (BOUNDARY) are specified relative
-        to the directory of the observation config.
+        File containing seismic observations (CSV or Parquet) and file containing the
+        polygon limiting observations used in the match step (BOUNDARY) are specified
+        relative to the directory of the observation config.
 
         Args:
             directory: Directory where observation config is located.
@@ -1088,6 +1088,13 @@ def from_obs_dict(
         name = ""
         filepath: str | Path | None = None
         boundary_filepath: str | Path | None = None
+
+        if "CSV" in observation_dict and "OBS_FILE" in observation_dict:
+            raise ObservationConfigError.with_context(
+                "SEISMIC_OBSERVATION cannot contain both 'CSV' and 'OBS_FILE'.",
+                observation_dict.context,
+            )
+
         for key, value in observation_dict.items():
             match key:
                 case "type":
@@ -1097,7 +1104,8 @@ def from_obs_dict(
                 case "CSV":
                     ConfigWarning.warn(
                         "CSV key is deprecated for seismic observations. "
-                        "Use OBS_FILE instead. "
+                        "Use OBS_FILE instead.",
+                        observation_dict.context,
                     )
                     filepath = value
                 case "OBS_FILE":
@@ -1107,19 +1115,13 @@ def from_obs_dict(
                 case _:
                     raise _unknown_key_error(str(key), observation_dict.context)
 
-        if "CSV" in observation_dict and "OBS_FILE" in observation_dict:
-            raise ObservationConfigError.with_context(
-                "SEISMIC_OBSERVATION cannot contain both 'CSV' and 'OBS_FILE'.",
-                observation_dict.context,
-            )
-
         if filepath is None:
             raise _missing_value_error(observation_dict.context, "OBS_FILE")
 
         filepath = Path(directory) / filepath
         if not filepath.exists():
             raise ObservationConfigError.with_context(
-                f"The CSV file ({filepath.absolute()}) "
+                f"The seismic observations file ({filepath.absolute()}) "
                 "does not exist or is not accessible.",
                 filepath,
             )

tests/ert/defaults_generator.py

@@ -3,6 +3,7 @@
 
 import numpy as np
 import polars as pl
+from lark import Token
 
 from ert.config._observations import (
     BreakthroughObservation,
@@ -11,7 +12,8 @@
     SeismicObservation,
     SummaryObservation,
 )
-from ert.config.parsing.observations_parser import ObservationType
+from ert.config.parsing.file_context_token import FileContextToken
+from ert.config.parsing.observations_parser import ObservationDict, ObservationType
 from ert.config.seismic_config import SeismicConfig
 
 
@@ -208,13 +210,23 @@ def create_seismic_observation_dict(
     name: str = "seismic_observation",
     csv: str = "horizon--amplitude_full_min_depth--20250101_20240101.csv",
     obs_file: str | None = None,
-) -> dict:
-    result = {"type": ObservationType.SEISMIC, "name": name}
+) -> ObservationDict:
+    data: dict = {"type": ObservationType.SEISMIC, "name": name}
     if obs_file is not None:
-        result["OBS_FILE"] = obs_file
+        data["OBS_FILE"] = obs_file
     else:
-        result["CSV"] = csv
-    return result
+        data["CSV"] = csv
+    context = FileContextToken(
+        Token(
+            type="foo",
+            line=2,
+            column=5,
+            end_column=13,
+            value="SEISMIC_OBSERVATION",
+        ),
+        "observations.txt",
+    )
+    return ObservationDict(data, context=context)
 
 
 def create_seismic_response(

tests/ert/unit_tests/config/test_observation_declaration.py

@@ -884,7 +884,7 @@ def test_that_non_existent_seismic_observation_file_raises_error(file_context_to
     )
 
 
-def test_that_missing_seismic_csv_filename_raises_error(file_context_token):
+def test_that_missing_seismic_observation_filename_raises_error(file_context_token):
     with pytest.raises(ObservationConfigError) as err:
         make_observations(
             "",
@@ -971,7 +971,13 @@ def test_that_missing_columns_in_seismic_parquet_observation_file_raises(
         }
     ).write_parquet("seismic_observations.parquet")
 
-    with pytest.raises(ObservationConfigError) as err:
+    with pytest.raises(
+        ObservationConfigError,
+        match=(
+            r"The seismic observations file seismic_observations.parquet "
+            r"is missing required column\(s\) X_UTME, Y_UTMN."
+        ),
+    ):
         make_observations(
             "",
             [
@@ -986,10 +992,6 @@ def test_that_missing_columns_in_seismic_parquet_observation_file_raises(
             ],
             shape_registry=ShapeRegistry(),
         )
-    assert (
-        "The seismic observations file seismic_observations.parquet "
-        "is missing required column(s) X_UTME, Y_UTMN." in str(err.value)
-    )
 
 
 @pytest.mark.usefixtures("use_tmpdir")

tests/ert/unit_tests/config/test_observations.py

@@ -2466,8 +2466,9 @@ def test_that_seismic_observation_dataframes_are_created_deprecated_csv_key(
     )
 
 
-def test_that_seismic_observation_dataframes_are_created_from_parquet_file(
-    mocked_files, file_context_token
+@pytest.mark.parametrize("file_format", ["parquet", "csv"])
+def test_that_seismic_observation_dataframes_are_created_from_obs_file(
+    mocked_files, file_context_token, file_format
 ):
     frame1 = pl.DataFrame(
         {
@@ -2488,11 +2489,16 @@ def test_that_seismic_observation_dataframes_are_created_from_parquet_file(
         }
     )
 
-    buf1, buf2 = BytesIO(), BytesIO()
-    frame1.write_parquet(buf1)
-    frame2.write_parquet(buf2)
-    mocked_files["obs1.parquet"] = buf1.getvalue()
-    mocked_files["obs2.parquet"] = buf2.getvalue()
+    if file_format == "parquet":
+        buf1, buf2 = BytesIO(), BytesIO()
+        frame1.write_parquet(buf1)
+        frame2.write_parquet(buf2)
+        mocked_files["obs1.parquet"] = buf1.getvalue()
+        mocked_files["obs2.parquet"] = buf2.getvalue()
+    else:
+        mocked_files["obs1.csv"] = frame1.write_csv()
+        mocked_files["obs2.csv"] = frame2.write_csv()
+
     ert_config = ErtConfig.from_dict(
         {
             "OBS_CONFIG": (
@@ -2502,15 +2508,15 @@ def test_that_seismic_observation_dataframes_are_created_from_parquet_file(
                         {
                             "type": ObservationType.SEISMIC,
                             "name": "NAME1",
-                            "OBS_FILE": "obs1.parquet",
+                            "OBS_FILE": f"obs1.{file_format}",
                         },
                         context=file_context_token(obs_type="SEISMIC_OBSERVATION"),
                     ),
                     ObservationDict(
                         {
                             "type": ObservationType.SEISMIC,
                             "name": None,
-                            "OBS_FILE": "obs2.parquet",
+                            "OBS_FILE": f"obs2.{file_format}",
                         },
                         context=file_context_token(obs_type="SEISMIC_OBSERVATION"),
                     ),