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/translator/dmptranslator.py

@@ -22,6 +22,7 @@ from saqc.lib.types import MaterializedGraph
 from saqc.core.translator.basetranslator import Translator, ForwardMap
 
 
+
 class DmpTranslator(Translator):
 
     """
@@ -38,6 +39,18 @@ class DmpTranslator(Translator):
         "BAD": BAD,
     }
 
+    _QUALITY_CAUSES = {
+        "BATTERY_LOW",
+        "BELOW_MINIMUM",
+        "ABOVE_MAXIMUM",
+        "BELOW_OR_ABOVE_MIN_MAX",
+        "ISOLATED_SPIKE",
+        "DEFECTIVE_SENSOR",
+        "LEFT_CENSORED_DATA",
+        "RIGHT_CENSORED_DATA",
+        "OTHER",
+    }
+
     def __init__(self):
         super().__init__(forward=self._FORWARD)
 
@@ -185,10 +198,18 @@ class DmpTranslator(Translator):
                 causes.append(cause)
                 comments.append(comment)
 
+            # DMP quality_cause needs some special care as only certain values
+            # and combinations are allowed.
+            # See: https://wiki.intranet.ufz.de/wiki/dmp/index.php/Qualit%C3%A4tsflags
+            causes = pd.Series(causes, index=flags[field].index)
+            causes[(causes == self.ARGUMENTS["cause"]) & (flags[field] > GOOD)] = "OTHER"
+            if not ((causes == "") | causes.isin(self._QUALITY_CAUSES)).all():
+                raise ValueError(f"quality causes needs to be one of {self._QUALITY_CAUSES}")
+
             var_flags = {
                 "quality_flag": tflags[field],
                 "quality_comment": pd.Series(comments, index=flags[field].index),
-                "quality_cause": pd.Series(causes, index=flags[field].index),
+                "quality_cause": causes,
             }
             out[field] = pd.DataFrame(var_flags)
         return pd.concat(out, axis="columns")