- Remove parameters extra names from avoiding collisions (smooth etc.)
- Optionally apply noise to generated data
- Add some datasets to generate_data from other libraries
- Separate error for local file and web file not found.
- Add examples to functions docstrings
- Add pd.Series to preprocessing as input.
- Generate sequentions even if bigger output than input.
- Remove create_inputs. Do one column input and data type in predictit.
- Add shape and dtype to numpy array type hints.
- Add features use_last_line (bool, optional): Last value may have not been complete yet. If False, it's removed. Defaults to False. reversit (bool, optional): If want to limit number of loaded lines, you can use select top and use reversed order. If you need to reverse it back, set this to True. Defaults to False
- Add Influx DB, Sqlite and postgres to databases
- Join load and consolidate to get_data
- Rename to mydataprocessing - load data and data consolidation to own submodule - get_data
- Define inplace argument in consolidation, data_preprocessing, and do inplace operations where possible
- In data consolidation, return also full scaler for all columns for inverse all columns and return in class maybe
- Feature extraction
- Finish generate test data modul -rename to datasets - add some real data, ramp etc...
- Add unit test for inputs (from visual) and generation
- In remove_outliers function do not remove it, but interpolate by neighbors option
- Check if somewhere in pandas iat instead of iloc (do it also for predictit)
- Add modin (or ray or dask) performance support