chore: validate results df has more than thresholded fields before up…

…loading (#679)

* validate results df has more than thresholded fields before uploading

* fix name in test

kolena/dataset/_common.py

@@ -84,6 +84,11 @@ def validate_dataframe_ids(df: pd.DataFrame, id_fields: List[str]) -> None:
     _validate_dataframe_ids_uniqueness(df, id_fields)
 
 
-def validate_dataframe_have_other_columns_besides_ids(df: pd.DataFrame, id_fields: List[str]) -> None:
-    if set(df.columns) == set(id_fields):
-        raise InputValidationError("dataframe only contains id fields")
+def validate_dataframe_columns(
+    df: pd.DataFrame,
+    id_fields: List[str],
+    thresholded_fields: Optional[List[str]] = None,
+) -> None:
+    minimal_fields = set(id_fields).union(thresholded_fields) if thresholded_fields else set(id_fields)
+    if set(df.columns) == minimal_fields:
+        raise InputValidationError("dataframe only contains id fields and thresholded fields")

kolena/dataset/evaluation.py

@@ -47,7 +47,7 @@
 from kolena.dataset._common import COL_THRESHOLDED_OBJECT
 from kolena.dataset._common import DEFAULT_SOURCES
 from kolena.dataset._common import validate_batch_size
-from kolena.dataset._common import validate_dataframe_have_other_columns_besides_ids
+from kolena.dataset._common import validate_dataframe_columns
 from kolena.dataset._common import validate_dataframe_ids
 from kolena.dataset.dataset import _load_dataset_metadata
 from kolena.dataset.dataset import _to_deserialized_dataframe
@@ -256,7 +256,7 @@ def _prepare_upload_results_request(
         if isinstance(df_result_input, pd.DataFrame):
             total_rows += df_result_input.shape[0]
             validate_dataframe_ids(df_result_input, id_fields)
-            validate_dataframe_have_other_columns_besides_ids(df_result_input, id_fields)
+            validate_dataframe_columns(df_result_input, id_fields, thresholded_fields)
             df_results = _process_result(config, df_result_input, id_fields, thresholded_fields)
             upload_data_frame(df=df_results, load_uuid=load_uuid)
 
@@ -265,7 +265,7 @@ def _prepare_upload_results_request(
             for df_result in df_result_input:
                 if not id_column_validated:
                     validate_dataframe_ids(df_result, id_fields)
-                    validate_dataframe_have_other_columns_besides_ids(df_result, id_fields)
+                    validate_dataframe_columns(df_result, id_fields, thresholded_fields)
                     id_column_validated = True
                 total_rows += df_result.shape[0]
                 df_results = _process_result(config, df_result, id_fields, thresholded_fields)

tests/integration/dataset/test_evaluation.py

@@ -25,6 +25,7 @@
 from kolena.dataset.dataset import _load_dataset_metadata
 from kolena.dataset.evaluation import _upload_results
 from kolena.errors import IncorrectUsageError
+from kolena.errors import InputValidationError
 from kolena.errors import NotFoundError
 from tests.integration.dataset.test_dataset import batch_iterator
 from tests.integration.helper import assert_frame_equal
@@ -490,6 +491,31 @@ def test__upload_results__thresholded() -> None:
     assert_frame_equal(fetched_df_result, expected_df_result, result_columns)
 
 
+def test__upload_results__only_id_and_thresholded_columns() -> None:
+    dataset_name = with_test_prefix(f"{__file__}::test__upload_results__only_id_and_thresholded_columns")
+    model_name = with_test_prefix(f"{__file__}::test__upload_results__only_id_and_thresholded_columns")
+    df_dp = get_df_dp()
+    dp_columns = [JOIN_COLUMN, "locator", "width", "height", "city"]
+    upload_dataset(dataset_name, df_dp[3:10][dp_columns], id_fields=ID_FIELDS)
+
+    records = [
+        dict(
+            user_dp_id=i,
+            bev=[dict(threshold=(j + 1) * 0.1, label="cat", foo=i + j) for j in range(3)],
+        )
+        for i in range(20)
+    ]
+    df_result = pd.DataFrame(records)
+
+    with pytest.raises(InputValidationError):
+        _upload_results(
+            dataset_name,
+            model_name,
+            df_result,
+            thresholded_fields=["bev"],
+        )
+
+
 def test__download_results__dataset_does_not_exist() -> None:
     dataset_name = with_test_prefix(f"{__file__}::test__download_results__dataset_does_not_exist")
     model_name = with_test_prefix(f"{__file__}::test__download_results__dataset_does_not_exist")

tests/unit/dataset/test_common.py

@@ -12,11 +12,12 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 from typing import List
+from typing import Optional
 
 import pandas as pd
 import pytest
 
-from kolena.dataset._common import validate_dataframe_have_other_columns_besides_ids
+from kolena.dataset._common import validate_dataframe_columns
 from kolena.dataset._common import validate_dataframe_ids
 from kolena.dataset._common import validate_id_fields
 from kolena.errors import DuplicateDatapointIdError
@@ -101,23 +102,38 @@ def test__validate_dataframe_ids__duplicate_id() -> None:
 
 
 @pytest.mark.parametrize(
-    "df, id_fields",
+    "df, id_fields, thresholded_fields",
     [
-        (pd.DataFrame(dict(a=[1, 2, 3], b=[1, 2, 1])), ["a"]),
-        (pd.DataFrame({"a.text": [1, 2, 3], "b.text": [1, 2, 1]}), ["a.text"]),
+        (pd.DataFrame(dict(a=[1, 2, 3], b=[1, 2, 1])), ["a"], None),
+        (pd.DataFrame({"a.text": [1, 2, 3], "b.text": [1, 2, 1]}), ["a.text"], None),
+        (pd.DataFrame({"a.text": [1, 2, 3], "b.text": [1, 2, 1]}), ["a.text"], []),
+        (
+            pd.DataFrame({"a.text": [1, 2, 3], "b.text": [1, 2, 1], "threshold": [0.1, 0.2, 0.3]}),
+            ["a.text"],
+            ["threshold"],
+        ),
     ],
 )
-def test__validate_dataframe_have_other_columns_besides_ids(df: pd.DataFrame, id_fields: List[str]) -> None:
-    validate_dataframe_have_other_columns_besides_ids(df, id_fields)
+def test__validate_dataframe_columns_besides_ids(
+    df: pd.DataFrame,
+    id_fields: List[str],
+    thresholded_fields: Optional[List[str]],
+) -> None:
+    validate_dataframe_columns(df, id_fields, thresholded_fields)
 
 
 @pytest.mark.parametrize(
-    "df, id_fields",
+    "df, id_fields, thresholded_fields",
     [
-        (pd.DataFrame(dict(a=[1, 2, 3], b=[1, 2, 1])), ["a", "b"]),
-        (pd.DataFrame({"a.text": [1, 2, 3], "b.text": [1, 2, 1]}), ["a.text", "b.text"]),
+        (pd.DataFrame(dict(a=[1, 2, 3], b=[1, 2, 1])), ["a", "b"], None),
+        (pd.DataFrame({"a.text": [1, 2, 3], "b.text": [1, 2, 1]}), ["a.text", "b.text"], None),
+        (pd.DataFrame({"a.text": [1, 2, 3], "b.text": [1, 2, 1]}), ["a.text"], ["b.text"]),
     ],
 )
-def test__validate_dataframe_have_other_columns_besides_ids__error(df: pd.DataFrame, id_fields: List[str]) -> None:
+def test__validate_dataframe_columns__error(
+    df: pd.DataFrame,
+    id_fields: List[str],
+    thresholded_fields: Optional[List[str]],
+) -> None:
     with pytest.raises(InputValidationError):
-        validate_dataframe_have_other_columns_besides_ids(df, id_fields)
+        validate_dataframe_columns(df, id_fields, thresholded_fields)