Update airfoil URL (#231)

* Update airfoil URL

* fix columns headers

Author: robsdavis

tests/plugins/core/models/helpers.py

@@ -4,9 +4,8 @@
 
 def get_airfoil_dataset() -> pd.DataFrame:
     df = pd.read_csv(
-        "https://www.neuraldesigner.com/files/datasets/airfoil_self_noise.csv",
-        # "https://archive.ics.uci.edu/static/public/291/airfoil+self+noise.zip", # TODO: change to this source?
-        sep=";",
+        "https://archive.ics.uci.edu/static/public/291/airfoil+self+noise.zip",
+        sep="\t",
         engine="python",
     )
     df.columns = df.columns.astype(str)

tests/plugins/domain_adaptation/da_helpers.py

@@ -25,9 +25,8 @@ def from_serde() -> Plugin:
 
 def get_airfoil_dataset() -> pd.DataFrame:
     df = pd.read_csv(
-        "https://www.neuraldesigner.com/files/datasets/airfoil_self_noise.csv",
-        # "https://archive.ics.uci.edu/static/public/291/airfoil+self+noise.zip", # TODO: change to this source?
-        sep=";",
+        "https://archive.ics.uci.edu/static/public/291/airfoil+self+noise.zip",
+        sep="\t",
         engine="python",
     )
     df.columns = df.columns.astype(str)

tests/plugins/generic/generic_helpers.py

@@ -31,9 +31,8 @@ def from_serde() -> Plugin:
 
 def get_airfoil_dataset() -> pd.DataFrame:
     df = pd.read_csv(
-        "https://www.neuraldesigner.com/files/datasets/airfoil_self_noise.csv",
-        # "https://archive.ics.uci.edu/static/public/291/airfoil+self+noise.zip", # TODO: change to this source?
-        sep=";",
+        "https://archive.ics.uci.edu/static/public/291/airfoil+self+noise.zip",
+        sep="\t",
         engine="python",
     )
     df.columns = df.columns.astype(str)

tests/plugins/privacy/fhelpers.py

@@ -25,9 +25,8 @@ def from_serde() -> Plugin:
 
 def get_airfoil_dataset() -> pd.DataFrame:
     df = pd.read_csv(
-        "https://www.neuraldesigner.com/files/datasets/airfoil_self_noise.csv",
-        # "https://archive.ics.uci.edu/static/public/291/airfoil+self+noise.zip", # TODO: change to this source?
-        sep=";",
+        "https://archive.ics.uci.edu/static/public/291/airfoil+self+noise.zip",
+        sep="\t",
         engine="python",
     )
     df.columns = df.columns.astype(str)

tests/utils/test_compression.py

@@ -8,10 +8,18 @@
 
 def get_airfoil_dataset() -> pd.DataFrame:
     df = pd.read_csv(
-        "https://www.neuraldesigner.com/files/datasets/airfoil_self_noise.csv",
-        # "https://archive.ics.uci.edu/static/public/291/airfoil+self+noise.zip", # TODO: change to this source?
-        sep=";",
+        "https://archive.ics.uci.edu/static/public/291/airfoil+self+noise.zip",
+        sep="\t",
         engine="python",
+        header=None,
+        names=[
+            "frequency",
+            "angle_of_attack",
+            "chord_length",
+            "free_stream_velocity",
+            "suction_side_displacement_thickness",
+            "scaled_sound_pressure_level",
+        ],
     )
     df.columns = df.columns.astype(str)
 
@@ -77,10 +85,7 @@ def test_compression_sanity2() -> None:
 def test_compression_sanity_airfoil() -> None:
     df = get_airfoil_dataset()
     df["chord_length"] = df["chord_length"].astype(str)
-    print(df.head())
     compressed_df, context = compress_dataset(df)
-    print(compressed_df)
-    print(context)
 
     assert len(compressed_df) == len(df)
     assert compressed_df.shape[1] > 0