separate out rename and replace

Author: betaBison

gnss_lib_py/parsers/navdata.py

@@ -51,7 +51,7 @@ def __init__(self, csv_path=None, pandas_df=None, numpy_array=None,
         else:
             self.build_navdata()
 
-        self.rename(rows=self._row_map(), inplace=True)
+        self.rename(self._row_map(), inplace=True)
 
         self.postprocess()
 
@@ -333,12 +333,10 @@ def __getitem__(self, key_idx):
                 "Cannot assign/return combination of strings and numbers"
         if np.all(row_str):
             # Return sliced strings
-            # arr_slice  = []
             arr_slice = np.atleast_2d(np.empty_like(self.array[rows, cols], dtype=object))
             for row_num, row in enumerate(row_list):
                 str_arr = self.get_strings(self.inv_map[row])
                 arr_slice[row_num, :] = str_arr[cols]
-                # arr_slice.append(str_arr[ cols])
         else:
             arr_slice = self.array[rows, cols]
 
@@ -420,7 +418,6 @@ def __setitem__(self, key_idx, new_value):
                         "String assignment only supported for ndarray of type object"
                 inv_map = self.inv_map
                 new_value = np.atleast_2d(new_value)
-                # new_value = np.reshape(new_value, [-1, new_value.shape[0]])
                 new_str_vals = np.ones_like(new_value, dtype=self.arr_dtype)
                 for row_num, row in enumerate(row_list):
                     key = inv_map[row]
@@ -755,8 +752,57 @@ def inv_map(self):
         inv_map = {v: k for k, v in self.map.items()}
         return inv_map
 
-    def rename(self, mapper=None, rows=None, inplace=False):
-        """Rename data within rows or row names of NavData class.
+    def rename(self, mapper=None, inplace=False):
+        """Rename rows of NavData class.
+
+        Row names must be strings.
+
+        Parameters
+        ----------
+        mapper : dict
+            Pairs of {"old_name" : "new_name"} for each row to be
+            renamed.
+        inplace : bool
+            If False, will return new NavData instance with rows
+            renamed. If True, will rename data rows in the
+            current NavData instance.
+
+        Returns
+        -------
+        new_navdata : gnss_lib_py.parsers.navdata.NavData or None
+            If inplace is False, returns NavData instance after renaming
+            specified rows. If inplace is True, returns
+            None.
+
+        """
+
+        if not isinstance(mapper, dict):
+            raise TypeError("'mapper' must be dict")
+        if not isinstance(inplace, bool):
+            raise TypeError("'inplace' must be bool")
+        for old_name in mapper:
+            if old_name not in self.map:
+                raise KeyError("'" + str(old_name) + "' row name " \
+                             + "doesn't exist in NavData class")
+
+        if not inplace:
+            new_navdata = self.copy()   # create copy to return
+        for old_name, new_name in mapper.items():
+            if not isinstance(new_name, str):
+                raise TypeError("New row names must be strings")
+            if inplace:
+                self.map[new_name] = self.map.pop(old_name)
+                self.str_map[new_name] = self.str_map.pop(old_name)
+            else:
+                new_navdata.map[new_name] = new_navdata.map.pop(old_name)
+                new_navdata.str_map[new_name] = new_navdata.str_map.pop(old_name)
+
+        if inplace:
+            return None
+        return new_navdata
+
+    def replace(self, mapper=None, rows=None, inplace=False):
+        """Replace data within rows or row names of NavData class.
 
         Row names must be strings.
 
@@ -772,23 +818,21 @@ def rename(self, mapper=None, rows=None, inplace=False):
             according to pairs of {"old_name" : "new_name"}.
             If mapper is not None, then an array-like input may be
             passed to indicate which rows of values should be remapped.
-
         inplace : bool
-            If False, will return new NavData instance with data and/or
-            rows renamed. If True, will rename data and/or rows in the
-            current NavData instance.
+            If False, will return new NavData instance with data
+            replaced. If True, will replace data in the current NavData
+            instance.
 
         Returns
         -------
         new_navdata : gnss_lib_py.parsers.navdata.NavData or None
-            If inplace is False, returns NavData instance after renaming
-            specified data and/or rows. If inplace is True, returns
-            None.
+            If inplace is False, returns NavData instance after
+            replacing specified data. If inplace is True, returns None.
 
         """
 
-        if not (isinstance(mapper, dict) or mapper is None):
-            raise TypeError("'mapper' must be dict or None")
+        if not isinstance(mapper, dict):
+            raise TypeError("'mapper' must be dict")
         if isinstance(rows,str):
             rows = [rows]
         if not (type(rows) in (dict, list, np.ndarray, tuple, set) \
@@ -806,7 +850,7 @@ def rename(self, mapper=None, rows=None, inplace=False):
             rows = None if len(rows)==0 else rows
         if not inplace:
             new_navdata = self.copy()   # create copy to return
-        if mapper is not None:
+        if mapper is not None and len(self) > 0:
             remap_rows = self.rows if rows is None else rows
             for row in remap_rows:
                 new_row_values = list(self[row])
@@ -816,16 +860,6 @@ def rename(self, mapper=None, rows=None, inplace=False):
                     self[row] = np.array(new_row_values)
                 else:
                     new_navdata[row] = np.array(new_row_values)
-        if isinstance(rows,dict):
-            for old_name, new_name in rows.items():
-                if not isinstance(new_name, str):
-                    raise TypeError("Row names must be strings")
-                if inplace:
-                    self.map[new_name] = self.map.pop(old_name)
-                    self.str_map[new_name] = self.str_map.pop(old_name)
-                else:
-                    new_navdata.map[new_name] = new_navdata.map.pop(old_name)
-                    new_navdata.str_map[new_name] = new_navdata.str_map.pop(old_name)
 
         if inplace:
             return None
@@ -857,8 +891,6 @@ def copy(self, rows=None, cols=None):
         new_navdata = NavData()
         inv_map = self.inv_map
         if rows is None:
-            # row_indices = slice(None, None).indices(len(self.rows))
-            # rows = np.arange(row_indices[0], row_indices[1], row_indices[2])
             rows = self.rows
         if cols is None:
             col_indices = slice(None, None).indices(len(self))

tests/parsers/test_navdata.py

@@ -371,14 +371,14 @@ def test_rename_inplace(pandas_df):
     data = NavData(pandas_df=pandas_df)
     data_temp = data.copy()
 
-    data_temp.rename(rows={"names": "terms"}, inplace=True)
+    data_temp.rename({"names": "terms"}, inplace=True)
     assert "names" not in data_temp.map
     assert "names" not in data_temp.str_map
     assert "terms" in data_temp.map
     assert "terms" in data_temp.str_map
 
     data_temp = data.copy()
-    data_temp.rename(rows={"floats": "decimals", "integers": "numbers"},
+    data_temp.rename(mapper={"floats": "decimals", "integers": "numbers"},
                 inplace = True)
     assert "floats" not in data_temp.map
     assert "floats" not in data_temp.str_map
@@ -392,15 +392,15 @@ def test_rename_inplace(pandas_df):
     # raises exception if input is not string
     data_temp = data.copy()
     with pytest.raises(TypeError):
-        data_temp.rename(rows={"names": 0}, inplace=True)
+        data_temp.rename({"names": 0}, inplace=True)
     data_temp = data.copy()
     with pytest.raises(TypeError):
-        data_temp.rename(rows={"names": 0.8}, inplace=True)
+        data_temp.rename({"names": 0.8}, inplace=True)
 
     # should raise error if key doesn't exist
     data_temp = data.copy()
     with pytest.raises(KeyError):
-        data_temp.rename(rows={"food": "test"}, inplace=True)
+        data_temp.rename({"food": "test"}, inplace=True)
 
 @pytest.mark.parametrize('pandas_df',
                         [
@@ -417,7 +417,7 @@ def test_rename_new_navdata(pandas_df):
     """
     data = NavData(pandas_df=pandas_df)
 
-    new_navdata = data.rename(rows={"names": "terms"})
+    new_navdata = data.rename({"names": "terms"})
     assert "names" not in new_navdata.map
     assert "names" not in new_navdata.str_map
     assert "terms" in new_navdata.map
@@ -428,7 +428,7 @@ def test_rename_new_navdata(pandas_df):
     assert "terms" not in data.map
     assert "terms" not in data.str_map
 
-    navdata = data.rename(rows={"floats": "decimals", "integers": "numbers"})
+    navdata = data.rename(mapper={"floats": "decimals", "integers": "numbers"})
     assert "floats" not in navdata.map
     assert "floats" not in navdata.str_map
     assert "integers" not in navdata.map
@@ -449,41 +449,75 @@ def test_rename_new_navdata(pandas_df):
 
     # raises exception if input is not string
     with pytest.raises(TypeError):
-        navdata = data.rename(rows={"names": 0})
+        navdata = data.rename({"names": 0})
     with pytest.raises(TypeError):
-        navdata = data.rename(rows={"names": 0.8})
+        navdata = data.rename({"names": 0.8})
 
     # should raise error if key doesn't exist
     with pytest.raises(KeyError):
-        navdata = data.rename(rows={"food": "test"})
+        navdata = data.rename({"food": "test"})
 
-@pytest.mark.parametrize('pandas_df',
-                        [
-                         lazy_fixture("df_simple"),
-                        ])
-def test_rename_fails(pandas_df):
-    """Test column renaming functionality.
+def test_replace_fails(df_simple, df_only_header):
+    """Test replace renaming functionality.
 
     Parameters
     ----------
-    pandas_df : pd.DataFrame
-        Dataframe for testing values
+    df_simple : pd.DataFrame
+        pd.DataFrame to initialize NavData with.
+    df_only_header : pd.DataFrame
+        Dataframe with only column names and no data
 
     """
-    data = NavData(pandas_df=pandas_df)
+    data = NavData(pandas_df=df_simple)
 
     with pytest.raises(TypeError) as excinfo:
-        data.rename(mapper=["names","floats"],rows={"names": "terms"})
+        data.replace(mapper=["names","floats"],rows={"names": "terms"})
     assert "mapper" in str(excinfo.value)
 
     with pytest.raises(TypeError) as excinfo:
-        data.rename(mapper=None,rows=1.0)
+        data.replace(mapper={"names":"terms"},rows=1.0)
     assert "rows" in str(excinfo.value)
 
     with pytest.raises(TypeError) as excinfo:
-        data.rename(rows={"names": "terms"}, inplace=0.)
+        data.replace({"names": "terms"}, inplace=0.)
+    assert "inplace" in str(excinfo.value)
+
+    # should raise error if key doesn't exist
+    with pytest.raises(KeyError):
+        data.replace({"gps":"GPS"},rows={"food": "test"}, inplace=True)
+
+    data = NavData(pandas_df=df_only_header)
+    data.replace({"gps":"GPS"},inplace=True)
+
+def test_rename_fails(df_simple, df_only_header):
+    """Test column renaming functionality.
+
+    Parameters
+    ----------
+    df_simple : pd.DataFrame
+        pd.DataFrame to initialize NavData with.
+    df_only_header : pd.DataFrame
+        Dataframe with only column names and no data
+
+    """
+    data = NavData(pandas_df=df_simple)
+
+    with pytest.raises(TypeError) as excinfo:
+        data.rename(mapper=["names","floats"])
+    assert "mapper" in str(excinfo.value)
+
+    with pytest.raises(TypeError) as excinfo:
+        data.rename(None)
+    assert "mapper" in str(excinfo.value)
+
+    with pytest.raises(TypeError) as excinfo:
+        data.rename(mapper={"names": "terms"}, inplace=0.)
     assert "inplace" in str(excinfo.value)
 
+    data = NavData(pandas_df=df_only_header)
+    data.rename({"strings":"words"},inplace=True)
+    assert "words" in data.rows
+    assert "strings" not in data.rows
 
 @pytest.mark.parametrize('rows',
                         [
@@ -505,7 +539,7 @@ def test_rename_fails(pandas_df):
                         np.array(["strings","integers","names","floats"]),
                         np.array(["integers","strings"]),
                         ])
-def test_rename_mapper_all(df_simple, rows):
+def test_replace_mapper_all(df_simple, rows):
     """Test data renaming functionality.
 
     Parameters
@@ -522,7 +556,7 @@ def test_rename_mapper_all(df_simple, rows):
               }
 
     # test that both rows change
-    new_navdata = data.rename(mapper, rows=rows)
+    new_navdata = data.replace(mapper, rows=rows)
     np.testing.assert_array_equal(new_navdata["strings"],
                           np.array(["GPS","glonass","galileo","GPS",
                                     "GPS","galileo"]))
@@ -536,7 +570,7 @@ def test_rename_mapper_all(df_simple, rows):
 
     # test that both rows change inplace
     data_temp = data.copy()
-    data_temp.rename(mapper, rows=rows, inplace=True)
+    data_temp.replace(mapper, rows=rows, inplace=True)
     np.testing.assert_array_equal(data_temp["strings"],
                           np.array(["GPS","glonass","galileo","GPS",
                                     "GPS","galileo"]))
@@ -564,7 +598,7 @@ def test_rename_mapper_all(df_simple, rows):
                         np.array(["floats","names","strings"]),
                         np.array(["strings"]),
                         ])
-def test_rename_mapper_partial(df_simple, rows):
+def test_replace_mapper_partial(df_simple, rows):
     """Test data renaming functionality.
 
     Parameters
@@ -581,7 +615,7 @@ def test_rename_mapper_partial(df_simple, rows):
               }
 
     # test that only "strings" changes and not "integers"
-    new_navdata = data.rename(mapper, rows=rows)
+    new_navdata = data.replace(mapper, rows=rows)
     np.testing.assert_array_equal(new_navdata["strings"],
                           np.array(["GPS","glonass","galileo","GPS",
                                     "GPS","galileo"]))
@@ -595,7 +629,7 @@ def test_rename_mapper_partial(df_simple, rows):
 
     # test that both rows change inplace
     data_temp = data.copy()
-    data_temp.rename(mapper, rows=rows, inplace=True)
+    data_temp.replace(mapper, rows=rows, inplace=True)
     np.testing.assert_array_equal(data_temp["strings"],
                           np.array(["GPS","glonass","galileo","GPS",
                                     "GPS","galileo"]))
@@ -608,7 +642,7 @@ def test_rename_mapper_partial(df_simple, rows):
     assert data_temp["strings"].dtype == object
 
 
-def test_rename_mapper_type_change(df_simple):
+def test_replace_mapper_type_change(df_simple):
     """Test data renaming functionality with type changes.
 
     Parameters
@@ -632,8 +666,8 @@ def test_rename_mapper_type_change(df_simple):
                       }
 
     # rename contents
-    data.rename(integers_mapper, inplace=True)
-    data.rename(strings_mapper, inplace=True)
+    data.replace(integers_mapper, inplace=True)
+    data.replace(strings_mapper, inplace=True)
 
     # make sure the rows hold the correct content
     np.testing.assert_array_equal(data["strings"],
@@ -670,7 +704,8 @@ def test_rename_mapper_and_rows(df_simple):
     row_mapper = {"integers" : "number_words"}
 
     # rename contents
-    data.rename(integers_mapper, rows=row_mapper, inplace=True)
+    data.replace(integers_mapper, rows="integers", inplace=True)
+    data.rename(row_mapper, inplace=True)
 
     # make sure the rows hold the correct content
     np.testing.assert_array_equal(data["number_words"],