fix(eda): remove unecessary compute

sfu-db · Sep 13, 2020 · 98c4ab0 · 98c4ab0
1 parent bad6a87
commit 98c4ab0
Show file tree

Hide file tree

Showing 3 changed files with 22 additions and 10 deletions.
diff --git a/dataprep/eda/missing/compute/bivariate.py b/dataprep/eda/missing/compute/bivariate.py
@@ -20,8 +20,7 @@
 from .common import LABELS, histogram
 
 
-@staged
-def compute_missing_bivariate(  # pylint: disable=too-many-locals
+def _compute_missing_bivariate(  # pylint: disable=too-many-locals
     df: DataArray,
     x: str,
     y: str,
@@ -32,7 +31,6 @@ def compute_missing_bivariate(  # pylint: disable=too-many-locals
     # pylint: disable=too-many-arguments
     """Calculate the distribution change on another column y when
     the missing values in x is dropped."""
-    df.compute("nulls")
 
     xloc = df.columns.get_loc(x)
     yloc = df.columns.get_loc(y)
@@ -144,3 +142,9 @@ def compute_missing_bivariate(  # pylint: disable=too-many-locals
             hist=df_ret, x=x, y=y, meta=meta["y"], visual_type="missing_impact_1v1",
         )
         return itmdt
+
+
+# Not using decorator here because jupyter autoreload does not support it.
+compute_missing_bivariate = staged(  # pylint: disable=invalid-name
+    _compute_missing_bivariate
+)
diff --git a/dataprep/eda/missing/compute/nullivariate.py b/dataprep/eda/missing/compute/nullivariate.py
@@ -15,8 +15,7 @@
 from ...staged import staged
 
 
-@staged
-def compute_missing_nullivariate(
+def _compute_missing_nullivariate(
     df: DataArray, bins: int
 ) -> Generator[Any, Any, Intermediate]:
     """Calculate the data for visualizing the plot_missing(df).
@@ -56,6 +55,12 @@ def compute_missing_nullivariate(
     )
 
 
+# Not using decorator here because jupyter autoreload does not support it.
+compute_missing_nullivariate = staged(  # pylint: disable=invalid-name
+    _compute_missing_nullivariate
+)
+
+
 def missing_perc_blockwise(bin_size: int) -> Callable[[np.ndarray], np.ndarray]:
     """Compute the missing percentage in a block."""
 

diff --git a/dataprep/eda/missing/compute/univariate.py b/dataprep/eda/missing/compute/univariate.py
@@ -19,14 +19,11 @@
 from .common import LABELS, histogram
 
 
-@staged
-def compute_missing_univariate(  # pylint: disable=too-many-locals
+def _compute_missing_univariate(  # pylint: disable=too-many-locals
     df: DataArray, x: str, bins: int, dtype: Optional[DTypeDef] = None,
 ) -> Generator[Any, Any, Intermediate]:
     """Calculate the distribution change on other columns when
     the missing values in x is dropped."""
-    df.compute("nulls")
-
     j = df.columns.get_loc(x)
 
     hists = {}
@@ -46,7 +43,7 @@ def compute_missing_univariate(  # pylint: disable=too-many-locals
             hist_range = (col0.min(axis=0), col0.max(axis=0))
 
         hists[col_name] = [
-            histogram(col, dtype=dtype, bins=bins, return_edges=True, range=hist_range,)
+            histogram(col, dtype=dtype, bins=bins, return_edges=True, range=hist_range)
             for col in [col0, col1]
         ]
 
@@ -97,3 +94,9 @@ def compute_missing_univariate(  # pylint: disable=too-many-locals
         dfs[col_name] = ret_df
 
     return Intermediate(data=dfs, x=x, meta=meta, visual_type="missing_impact_1vn")
+
+
+# Not using decorator here because jupyter autoreload does not support it.
+compute_missing_univariate = staged(  # pylint: disable=invalid-name
+    _compute_missing_univariate
+)