Follow-Up Translations

This merge requests incorporates the following changes/additions:

• Allow the use of translator specific keywords in function calls, e.g. comment and cause for the DmpTranslator
• Generate valid Historys from given input flags. The user flags usually lack information (i.e. concrete parameter values, order of function calls, ...) to infer a real history. I loosed the definition of the word valid, to 'rerunning the history inferred from a DataFrame of user flags on the same input data, needs to yield the same user flags again'
• Make the different translators available through the CLI (i.e. the exe)

saqc/core/history.py

@@ -78,7 +78,7 @@ class History:
             hist = hist.copy()
             mask = mask.copy()
 
-        self.hist = hist
+        self.hist = hist.astype("category")
         self.mask = mask
 
     @property
@@ -166,7 +166,7 @@ class History:
             touched = s.notna()
             self.mask.iloc[touched, :pos] = False
 
-        self.hist[pos] = s
+        self.hist[pos] = s.astype("category")
 
         return self
 
@@ -269,7 +269,7 @@ class History:
         # clear the current mask
         self.mask.loc[(~value_mask & value_hist.notna()).any(axis="columns")] = False
 
-        self.hist.loc[:, columns] = value_hist.copy()
+        self.hist.loc[:, columns] = value_hist.astype("category")
         self.mask.loc[:, columns] = value_mask.copy()
         return self
 
@@ -316,7 +316,7 @@ class History:
         # the last column maybe is not entirely True, but
         # the following append, will fix this
         self.hist = self.hist.iloc[:, :-n]
-        self.mask = self.mask.iloc[:, :-n]
+        self.mask = self.mask.iloc[:, :-n].astype("category")
 
         self.append(s)
         return self
@@ -383,10 +383,14 @@ class History:
         -------
         History
         """
-        self.hist = self.hist.reindex(index=index, copy=False, fill_value=np.nan)
+        self.hist = self.hist.reindex(
+            index=index, copy=False, fill_value=np.nan
+        ).astype("category")
         self.mask = self.mask.reindex(index=index, copy=False, fill_value=False)
         # Note: all following code must handle empty frames
-        self.hist.iloc[:, -1:] = self.hist.iloc[:, -1:].fillna(fill_value_last)
+        self.hist.iloc[:, -1:] = (
+            self.hist.iloc[:, -1:].fillna(fill_value_last).astype("category")
+        )
         self.mask.iloc[:, -1:] = True
         return self
 
@@ -467,8 +471,8 @@ class History:
                 f"'hist' must be of type pd.DataFrame, but {type(obj).__name__} was given"
             )
 
-        if (obj.dtypes != float).any():
-            raise ValueError("dtype of all columns in hist must be float")
+        # if (obj.dtypes != float).any():
+        #     raise ValueError("dtype of all columns in hist must be float")
 
         if not obj.empty and (
             not obj.columns.equals(pd.Index(range(len(obj.columns))))
@@ -490,8 +494,8 @@ class History:
                 f"value must be of type pd.Series, but {type(obj).__name__} was given"
             )
 
-        if not obj.dtype == float:
-            raise ValueError("dtype must be float")
+        # if not obj.dtype == float:
+        #     raise ValueError("dtype must be float")
 
         return obj
 
@@ -545,12 +549,16 @@ def applyFunctionOnHistory(
     new_history = History()
 
     if func_handle_df:
-        history.hist = hist_func(history.hist, **hist_kws)
+        history.hist = hist_func(history.hist.astype(float), **hist_kws).astype(
+            "category"
+        )
         history.mask = hist_func(history.mask, **mask_kws)
 
     else:
         for pos in history.columns:
-            new_history.hist[pos] = hist_func(history.hist[pos], **hist_kws)
+            new_history.hist[pos] = hist_func(
+                history.hist[pos].astype(float), **hist_kws
+            ).astype("category")
             new_history.mask[pos] = mask_func(history.mask[pos], **mask_kws)
 
     # handle unstable state
@@ -560,7 +568,7 @@ def applyFunctionOnHistory(
         if isinstance(last_column, str) and last_column == "dummy":
             last_column = pd.Series(UNTOUCHED, index=new_history.index, dtype=float)
 
-        new_history.append(last_column, force=True)
+        new_history.append(last_column.astype("category"), force=True)
 
     # assure a boolean mask and UNFLAGGED column
     new_history.mask = new_history.mask.fillna(True).astype(bool)