From dbdade94288c97d8e6842c7733e7ecd12c852b9b Mon Sep 17 00:00:00 2001
From: David Schaefer <david.schaefer@ufz.de>
Date: Thu, 27 Jun 2019 17:03:20 +0200
Subject: [PATCH] various modifactions
---
saqc/flagger/baseflagger.py | 29 ++++++++++++++++-------------
saqc/flagger/dmpflagger.py | 10 ++++------
2 files changed, 20 insertions(+), 19 deletions(-)
diff --git a/saqc/flagger/baseflagger.py b/saqc/flagger/baseflagger.py
index d2577cbd6..f367250b3 100644
--- a/saqc/flagger/baseflagger.py
+++ b/saqc/flagger/baseflagger.py
@@ -41,14 +41,14 @@ class BaseFlagger:
pandas data structures tend to behave unpredictively
in assignments, especially if a multi column index is used
"""
- if flag is None:
- flag = self.flags.max()
- else:
- self._checkFlag(flag)
+ flag = self.flags.max() if flag is None else self._checkFlag(flag)
flags = flags.copy()
- # NOTE: conversion of "flags frame to np.array is done here already, since return argument is just the array
- # anyway. For applying mulitdimensional indexing on the DataFrame "flags", you would have to stack it first,
- # with .stack() (try flags.stack[flags<flag] = flag and than unstack.)
+ # NOTE:
+ # conversion of 'flags' frame to np.array is done here already,
+ # since return argument is just the array anyway. For applying
+ # mulitdimensional indexing on the DataFrame 'flags', you would
+ # have to stack it first
+ # (try flags.stack[flags<flag] = flag and than unstack.)
flags = flags.values
flags[flags < flag] = flag
@@ -57,20 +57,23 @@ class BaseFlagger:
def initFlags(self, data: pd.DataFrame) -> pd.DataFrame:
out = data.copy()
out[:] = self.flags[0]
- # astype conversion of return Dataframe performed seperately, because pd.DataFrame(...,dtype=self.flags)
- # wont give you categorical flag objects:
+ # NOTE:
+ # astype conversion of return Dataframe performed
+ # seperately, because pd.DataFrame(..., dtype=self.flags)
+ # wont give you categorical flag objects
return out.astype(self.flags)
def isFlagged(self, flags: ArrayLike, flag: T = None) -> ArrayLike:
if flag is None:
return pd.notnull(flags) & (flags > self.flags[0])
- self._checkFlag(flag)
- return flags == flag
+ return flags == self._checkFlag(flag)
def _checkFlag(self, flag):
if flag not in self.flags:
- raise ValueError(f"Invalid flag '{flag}'. "
- f"Possible choices are {list(self.flags.categories)[1:]}")
+ raise ValueError(
+ f"Invalid flag '{flag}'. "
+ f"Possible choices are {list(self.flags.categories)[1:]}")
+ return flag
def nextTest(self):
pass
diff --git a/saqc/flagger/dmpflagger.py b/saqc/flagger/dmpflagger.py
index e1c6451cf..2ec2b2d44 100644
--- a/saqc/flagger/dmpflagger.py
+++ b/saqc/flagger/dmpflagger.py
@@ -36,10 +36,11 @@ class DmpFlagger(BaseFlagger):
self.project_version = version.decode().strip()
def initFlags(self, data, **kwargs):
- columns = data.columns if isinstance(data, pd.DataFrame) else [data.name]
+ if isinstance(data, pd.Series):
+ data = data.to_frame()
colindex = pd.MultiIndex.from_product(
- [columns, self.flag_fields],
+ [data.columns, self.flag_fields],
names=[ColumnLevels.VARIABLES, ColumnLevels.FLAGS])
out = pd.DataFrame(data=self.flags[0],
@@ -53,10 +54,7 @@ class DmpFlagger(BaseFlagger):
if not isinstance(flags, pd.DataFrame):
raise TypeError
- if flag is None:
- flag = self.flags.max()
- else:
- self._checkFlag(flag)
+ flag = self.flags.max() if flag is None else self._checkFlag(flag)
if Keywords.VERSION in comment:
comment = comment.replace(Keywords.VERSION, self.project_version)
--
GitLab