Merge pull request #35 from FAST-HEP/BK_implement_mapping_multiply

Implement multiply by mapping

CHANGELOG.md

@@ -4,6 +4,10 @@ All notable changes to this project will be documented in this file.
 The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/),
 and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
 
+## [0.6.5] - 2020-05-12
+### Added
+- Implement the multiply_values with a mapping, PR #35 [@benkrikler](https://github.com/benkrikler)
+
 ## [0.6.4] - 2020-05-07
 ### Added
 - New postprocessing stage to filter columns, PR #34 [@benkrikler](https://github.com/benkrikler)

fast_plotter/postproc/functions.py

@@ -382,17 +382,21 @@ def multiply_values(df, constant=0, mapping={}, weight_by_dataframes=[], apply_i
         if mask.dtype.kind != "b":
             msg = "'apply_if' statement doesn't return a boolean: %s"
             raise ValueError(msg % apply_if)
+        index = df.index
         ignored = df.loc[~mask]
         df = df.loc[mask]
     if mapping:
-        raise NotImplementedError("'mapping' option not yet implemented")
+        for select, value in mapping.items():
+            df = multiply_values(df, constant=value, apply_if=select)
     if weight_by_dataframes:
         for mul_df in weight_by_dataframes:
             df = multiply_dataframe(df, mul_df)
     if constant:
-        df = df * constant
+        numeric_cols = df.select_dtypes('number')
+        df = df.assign(**numeric_cols.multiply(constant))
     if apply_if:
-        df = pd.concat([df, ignored])
+        df = pd.concat([df, ignored]).reindex(index=index)
+
     return df
 
 

fast_plotter/version.py

@@ -12,5 +12,5 @@ def split_version(version):
     return tuple(result)
 
 
-__version__ = '0.6.4'
+__version__ = '0.6.5'
 version_info = split_version(__version__) # noqa

setup.cfg

@@ -1,5 +1,5 @@
 [bumpversion]
-current_version = 0.6.4
+current_version = 0.6.5
 commit = True
 tag = False
 

tests/postproc/test_functions.py

@@ -143,9 +143,24 @@ def test_filter_cols(binned_df):
 #     #def merge(dfs):
 #     pass
 
-# def test_multiply_values():
-#     #def multiply_values(df, constant=0, mapping={}, weight_by_dataframes=[], apply_if=None):
-#     pass
+def test_multiply_values(binned_df):
+    result = funcs.multiply_values(binned_df, constant=3)
+    assert np.array_equal(result.a, np.arange(len(binned_df)) * 3)
+    assert np.array_equal(result.b, binned_df.b)
+
+    result = funcs.multiply_values(binned_df, apply_if="int % 2 == 0", constant=3)
+    assert np.array_equal(result.a, np.arange(len(binned_df)) * np.repeat([3, 1, 3, 1], 10))
+    assert np.array_equal(result.b, binned_df.b)
+
+    result = funcs.multiply_values(binned_df, mapping={"int % 2 == 0": 3, "int % 2 == 1": 7.2})
+    assert np.array_equal(result.a, np.arange(len(binned_df)) * np.repeat([3, 7.2, 3, 7.2], 10))
+    assert np.array_equal(result.b, binned_df.b)
+
+    result = funcs.multiply_values(binned_df, mapping={"cat=='foo'": 1.2, "cat=='bar'": 19})
+    tiled_vals = np.tile(np.repeat([1.2, 19], 5), 4)
+    assert np.array_equal(result.a, np.arange(len(binned_df)) * tiled_vals)
+    assert np.array_equal(result.b, binned_df.b)
+
 
 # def test_multiply_dataframe():
 #     #def multiply_dataframe(df, multiply_df, use_column=None):