movielens WIP

Signed-off-by: miguelgfierro <miguelgfierro@users.noreply.github.com>

tests/ci/azureml_tests/test_groups.py

@@ -15,12 +15,12 @@
 
 nightly_test_groups = {
     "group_cpu_001": [  # Total group time: 1883s
-        "tests/smoke/recommenders/dataset/test_movielens.py::test_download_and_extract_movielens",  # 0.45s
+        "tests/data_validation/recommenders/datasets/test_movielens.py::test_download_and_extract_movielens",  # 0.45s + 0.61s + 3.47s + 8.28s
+        #
         "tests/smoke/recommenders/dataset/test_movielens.py::test_load_item_df",  # 0.47s
         "tests/smoke/recommenders/dataset/test_movielens.py::test_load_pandas_df",  # 2.45s
         #
         "tests/integration/recommenders/datasets/test_movielens.py::test_load_pandas_df",  # 16.87s
-        "tests/integration/recommenders/datasets/test_movielens.py::test_download_and_extract_movielens",  # 0.61s + 3.47s + 8.28s
         "tests/integration/recommenders/datasets/test_movielens.py::test_load_item_df",  # 0.59s + 3.59s + 8.44s
         "tests/integration/recommenders/datasets/test_movielens.py::test_load_pandas_df",  # 37.33s + 352.99s + 673.61s
         #
@@ -156,11 +156,13 @@
 
 pr_gate_test_groups = {
     "group_cpu_001": [  # Total group time: 525.96s
-        "tests/data_validation/recommenders/datasets/test_movielens.py::test_load_pandas_df_mock_100__with_default_param__succeed",
-        "tests/data_validation/recommenders/datasets/test_movielens.py::test_load_pandas_df_mock_100__with_custom_param__succeed",
         "tests/data_validation/recommenders/datasets/test_movielens.py::test_mock_movielens_schema__has_default_col_names",
-        "tests/data_validation/recommenders/datasets/test_movielens.py::test_mock_movielens_schema__get_df__return_success",
         "tests/data_validation/recommenders/datasets/test_movielens.py::test_mock_movielens_schema__get_df_remove_default_col__return_success",
+        "tests/data_validation/recommenders/datasets/test_movielens.py::test_mock_movielens_schema__get_df_invalid_param__return_failure",
+        "tests/data_validation/recommenders/datasets/test_movielens.py::test_mock_movielens_schema__get_df__return_success",
+        "tests/data_validation/recommenders/datasets/test_movielens.py::test_mock_movielens_data__no_name_collision",
+        "tests/data_validation/recommenders/datasets/test_movielens.py::test_load_pandas_df_mock_100__with_default_param__succeed",
+        "tests/data_validation/recommenders/datasets/test_movielens.py::test_load_pandas_df_mock_100__with_custom_param__succeed",
         "tests/data_validation/recommenders/datasets/test_wikidata.py::test_find_wikidata_id",
         "tests/unit/recommenders/datasets/test_download_utils.py::test_maybe_download_wrong_bytes",
         "tests/unit/recommenders/datasets/test_download_utils.py::test_maybe_download_maybe",
@@ -205,11 +207,11 @@
         "tests/unit/examples/test_notebooks_python.py::test_sar_single_node_runs",
     ],
     "group_spark_001": [  # Total group time: 270.41s
-        "tests/data_validation/recommenders/datasets/test_movielens.py::test_load_spark_df_mock_100__with_custom_param__succeed",
         "tests/data_validation/recommenders/datasets/test_movielens.py::test_mock_movielens_schema__get_spark_df__return_success",
-        "tests/data_validation/recommenders/datasets/test_movielens.py::test_mock_movielens_schema__get_spark_df__data_serialization_default_param",
         "tests/data_validation/recommenders/datasets/test_movielens.py::test_mock_movielens_schema__get_spark_df__store_tmp_file",
+        "tests/data_validation/recommenders/datasets/test_movielens.py::test_mock_movielens_schema__get_spark_df__data_serialization_default_param",
         "tests/data_validation/recommenders/datasets/test_movielens.py::test_load_spark_df_mock_100__with_default_param__succeed",
+        "tests/data_validation/recommenders/datasets/test_movielens.py::test_load_spark_df_mock_100__with_custom_param__succeed",
         "tests/unit/recommenders/datasets/test_spark_splitter.py::test_stratified_splitter",
         "tests/unit/recommenders/datasets/test_spark_splitter.py::test_chrono_splitter",
         "tests/unit/recommenders/evaluation/test_spark_evaluation.py::test_user_diversity_item_feature_vector",

tests/data_validation/recommenders/datasets/test_movielens.py

@@ -5,11 +5,27 @@
 import os
 import pandas
 from pandas.core.series import Series
-from pytest_mock import MockerFixture
 import pytest
+from pytest_mock import MockerFixture
+
+try:
+    from pyspark.sql.types import (
+        StructType,
+        StructField,
+        IntegerType,
+    )
+    from pyspark.sql.functions import col
+except ImportError:
+    pass  # skip this import if we are in pure python environment
 
 from recommenders.datasets.movielens import MockMovielensSchema
-from recommenders.datasets.movielens import load_pandas_df, load_spark_df
+from recommenders.datasets.movielens import (
+    load_pandas_df,
+    load_spark_df,
+    load_item_df,
+    download_movielens,
+    extract_movielens,
+)
 from recommenders.datasets.movielens import (
     DATA_FORMAT,
     MOCK_DATA_FORMAT,
@@ -66,6 +82,53 @@ def test_mock_movielens_schema__get_df__return_success(
         assert len(df[DEFAULT_GENRE_COL]) == size
 
 
+def test_mock_movielens_data__no_name_collision():
+    """
+    Making sure that no common names are shared between the mock and real dataset sizes
+    """
+    dataset_name = set(DATA_FORMAT.keys())
+    dataset_name_mock = set(MOCK_DATA_FORMAT.keys())
+    collision = dataset_name.intersection(dataset_name_mock)
+    assert not collision
+
+
+def test_load_pandas_df_mock_100__with_default_param__succeed():
+    df = load_pandas_df("mock100")
+    assert type(df) == pandas.DataFrame
+    assert len(df) == 100
+    assert not df[[DEFAULT_USER_COL, DEFAULT_ITEM_COL]].duplicated().any()
+
+
+def test_load_pandas_df_mock_100__with_custom_param__succeed():
+    df = load_pandas_df(
+        "mock100", title_col=DEFAULT_TITLE_COL, genres_col=DEFAULT_GENRE_COL
+    )
+    assert type(df[DEFAULT_TITLE_COL]) == Series
+    assert type(df[DEFAULT_GENRE_COL]) == Series
+    assert len(df) == 100
+    assert "|" in df.loc[0, DEFAULT_GENRE_COL]
+    assert df.loc[0, DEFAULT_TITLE_COL] == "foo"
+
+
+@pytest.mark.parametrize("size", ["100k", "1m", "10m", "20m"])
+def test_download_and_extract_movielens(size, tmp):
+    """Test movielens data download and extract"""
+    zip_path = os.path.join(tmp, "ml.zip")
+    download_movielens(size, dest_path=zip_path)
+    assert len(os.listdir(tmp)) == 1
+    assert os.path.exists(zip_path)
+
+    rating_path = os.path.join(tmp, "rating.dat")
+    item_path = os.path.join(tmp, "item.dat")
+    extract_movielens(
+        size, rating_path=rating_path, item_path=item_path, zip_path=zip_path
+    )
+    # Test if raw-zip file, rating file, and item file are cached
+    assert len(os.listdir(tmp)) == 3
+    assert os.path.exists(rating_path)
+    assert os.path.exists(item_path)
+
+
 @pytest.mark.spark
 @pytest.mark.parametrize("keep_genre_col", [True, False])
 @pytest.mark.parametrize("keep_title_col", [True, False])
@@ -109,29 +172,12 @@ def test_mock_movielens_schema__get_spark_df__data_serialization_default_param(
     assert df.count() == data_size
 
 
-def test_mock_movielens_data__no_name_collision():
-    """
-    Making sure that no common names are shared between the mock and real dataset sizes
-    """
-    dataset_name = set(DATA_FORMAT.keys())
-    dataset_name_mock = set(MOCK_DATA_FORMAT.keys())
-    collision = dataset_name.intersection(dataset_name_mock)
-    assert not collision
-
-
 @pytest.mark.spark
 def test_load_spark_df_mock_100__with_default_param__succeed(spark):
     df = load_spark_df(spark, "mock100")
     assert df.count() == 100
 
 
-def test_load_pandas_df_mock_100__with_default_param__succeed():
-    df = load_pandas_df("mock100")
-    assert type(df) == pandas.DataFrame
-    assert len(df) == 100
-    assert not df[[DEFAULT_USER_COL, DEFAULT_ITEM_COL]].duplicated().any()
-
-
 @pytest.mark.spark
 def test_load_spark_df_mock_100__with_custom_param__succeed(spark):
     df = load_spark_df(
@@ -142,14 +188,3 @@ def test_load_spark_df_mock_100__with_custom_param__succeed(spark):
     assert df.count() == 100
     assert "|" in df.take(1)[0][DEFAULT_GENRE_COL]
     assert df.take(1)[0][DEFAULT_TITLE_COL] == "foo"
-
-
-def test_load_pandas_df_mock_100__with_custom_param__succeed():
-    df = load_pandas_df(
-        "mock100", title_col=DEFAULT_TITLE_COL, genres_col=DEFAULT_GENRE_COL
-    )
-    assert type(df[DEFAULT_TITLE_COL]) == Series
-    assert type(df[DEFAULT_GENRE_COL]) == Series
-    assert len(df) == 100
-    assert "|" in df.loc[0, DEFAULT_GENRE_COL]
-    assert df.loc[0, DEFAULT_TITLE_COL] == "foo"

tests/integration/recommenders/datasets/test_movielens.py

@@ -271,23 +271,3 @@ def test_load_spark_df(
     assert len(df.columns) == 4
     del df
     gc.collect()
-
-
-@pytest.mark.integration
-@pytest.mark.parametrize("size", ["1m", "10m", "20m"])
-def test_download_and_extract_movielens(size, tmp):
-    """Test movielens data download and extract"""
-    zip_path = os.path.join(tmp, "ml.zip")
-    download_movielens(size, dest_path=zip_path)
-    assert len(os.listdir(tmp)) == 1
-    assert os.path.exists(zip_path)
-
-    rating_path = os.path.join(tmp, "rating.dat")
-    item_path = os.path.join(tmp, "item.dat")
-    extract_movielens(
-        size, rating_path=rating_path, item_path=item_path, zip_path=zip_path
-    )
-    # Test if raw-zip file, rating file, and item file are cached
-    assert len(os.listdir(tmp)) == 3
-    assert os.path.exists(rating_path)
-    assert os.path.exists(item_path)

tests/smoke/recommenders/dataset/test_movielens.py

@@ -11,16 +11,6 @@
     extract_movielens,
 )
 
-try:
-    from pyspark.sql.types import (
-        StructType,
-        StructField,
-        IntegerType,
-    )
-    from pyspark.sql.functions import col
-except ImportError:
-    pass  # skip this import if we are in pure python environment
-
 
 @pytest.mark.smoke
 @pytest.mark.parametrize(
@@ -210,23 +200,3 @@ def test_load_spark_df(
     assert df.count() == num_samples
     # user, item, rating and timestamp
     assert len(df.columns) == 4
-
-
-@pytest.mark.smoke
-@pytest.mark.parametrize("size", ["100k"])
-def test_download_and_extract_movielens(size, tmp):
-    """Test movielens data download and extract"""
-    zip_path = os.path.join(tmp, "ml.zip")
-    download_movielens(size, dest_path=zip_path)
-    assert len(os.listdir(tmp)) == 1
-    assert os.path.exists(zip_path)
-
-    rating_path = os.path.join(tmp, "rating.dat")
-    item_path = os.path.join(tmp, "item.dat")
-    extract_movielens(
-        size, rating_path=rating_path, item_path=item_path, zip_path=zip_path
-    )
-    # Test if raw-zip file, rating file, and item file are cached
-    assert len(os.listdir(tmp)) == 3
-    assert os.path.exists(rating_path)
-    assert os.path.exists(item_path)