Aethos

Hotfixed NLTK package in setup.py and the vulnerable version

Release Notes
---
Introducing visualizations

General
---
- Param for Transformers can now be passed as kwargs

Bugfixes
---
- Fixed bug where rejoining dataframes after a technique would fail due to index not being reset on split

Features
---
- Added the preprocessing splitting text into sentences

Visualizations
---
- Added box plot
- Added raincloud plot

Release Notes
---
- Test and train data is now only created from a full dataset when the user doesn't specify they are using the entire dataset.
- Can now reference column names using . notation from the clean, preprocess and feature object
- Can now assign values to the data, test_data and train_data attributes of each object
- Analytic techniques now return the entire object back so you can keep copies for different branches of analysis

Bugfixes
---
- Apply function now properly returns the entire dataset.
- Logs are no longer repeating when creating copies of the object

Features
---
- Added new preprocessing technique that splits documents into sentences using nltk

Release Notes
---
- Hotfixed bugs with reporting
- Touched up reporting so it's more verbose and better formatted

Release Notes
---

- Increased overall quality of life and usability of library
- Columns as input for functions can now be passed in as *args ('col1', 'col2', ..) or a list (['col1', 'col2', ..])
- Functions now will apply the analysis to the whole dataset or the columns that fit the type (i.e. if no columns are provided for replacing missing with mean values, it will apply the analysis on all numeric columns)
- Functions no longer return an output of the entire dataframe, they just now return the output of the first 10 rows of the full dataset or the training dataset to see the results of your analysis

Release Notes
--------------
Hotfixed the OnehotEncoder technique
Hotfixed the apply technique