- Aug 03, 2020
-
-
David Schäfer authored
Bump pyarrow from 0.16.0 to 1.0.0 See merge request !67
-
David Schäfer authored
it, matplotlib will choose one, according to the available libraries on the system
-
- Jul 31, 2020
-
-
WKDV Bot authored
Bumps [pyarrow](https://arrow.apache.org/) from 0.16.0 to 1.0.0.
-
- Jul 29, 2020
-
-
Peter Lünenschloß authored
-
- Jul 28, 2020
-
-
Bert Palm authored
a implementation with a finally clause would be possible, but make the code look like spaghetti. :P
-
- Jul 27, 2020
-
-
Bert Palm authored
i removed the setup of the logging module from the core. It interferes with the needs of the user of the lib (saqc). eg. Messages are double printed or not at all. >> It is strongly advised that you do not add any handlers other than >> NullHandler to your library’s loggers. [...] from https://docs.python.org/3/howto/logging.html we might add a defalutLoggingSetup(log_level='INFO') function or similar in future, to ease the setup the logging.
-
- Jul 22, 2020
- Jul 20, 2020
-
- Jul 17, 2020
-
-
Bert Palm authored
-
Bert Palm authored
Data that is flagger.BAD is masked by default. One can pass a flag or a list-like to the param to_mask to control the behavior. Pass an empty list to disable masking. Nevertheless IFF the index changed the masked data is not restored. Also flags are NOT restored IFF they considered _flagged_ before the func call, but not afterwards anymore. Also any data that is set on masked locations is not overwritten by unmasking.
-
Bert Palm authored
Now giving an empty list to isFlagged return the correct result, an all-False series/dios. flagger.isFlagged(flag=[]) Before a nearly all-True data was returned, because of a call to notna(), that was not protected by the loop.
-
Bert Palm authored
-
Bert Palm authored
-
- Jul 16, 2020
-
-
David Schäfer authored
-
- Jul 15, 2020
-
-
Peter Lünenschloß authored
Patternrec See merge request !50
-
Juliane Geller authored
-
Juliane Geller authored
-
Peter Lünenschloß authored
-
David Schäfer authored
-
-
David Schäfer authored
seems to be hard to stick to naming conventions
-
- Jul 14, 2020
-
-
Juliane Geller authored
-
Juliane Geller authored
-
- Jul 13, 2020
-
-
Juliane Geller authored
-
-
Juliane Geller authored
-
Juliane Geller authored
-
Juliane Geller authored
-
Juliane Geller authored
-
Juliane Geller authored
-
- Jul 12, 2020
-
-
David Schäfer authored
`write_back` parameter in `getResult`. While the previous version worked in the sense, that it allowed to `save` computations already done, I disliked the state-changing behaviour of that implemetation. Now we have a way to realize the registered test functions into a new `SaQC` without modifying the old one and don't violate the functional interface of the `SaQC`-Objects any more.
-
David Schäfer authored
the actually needed charactersitics. The old check was forcing equal columns on `flagger.flags` and `data`, a condition we don't need to be satisfied and don't satisfy during the lifetime of `SaQC`-Object either. The new check only guarantees, that all columns in `data` have a counterpart in `flagger.flags` but allow the latter to have additional columns as well.
-
David Schäfer authored
-
- Jul 11, 2020
-
-
David Schäfer authored
-
David Schäfer authored
-
Peter Lünenschloß authored
-
Peter Lünenschloß authored
-