Dbcollection

Changes

- 157 Update `cifar10` classification task code format
- 158 Update `cifar100` classification task code format
- 159 Update `mnist` classification task code format

Changes

- 139 144 Refactor `cifar10` dataset + docs
- 142 Refactor `cifar100` dataset + docs
- 141 Add unit tests for `mnist`

This release brings some changes and fixes to the core package, more specifically to the API. The bulk of the changes to the code is due to a major refactor of the internals of the package and unit tests. This release starts the process to refactor the dataset's task scripts to simplify coding the metadata processing of a task and to introduce unit tests to the scripts.

Relevant changes

- `config_cache()` was renamed to `cache()`
- info methods were merged into a single one (111)

This release brings a new feature to `get()` with the `convert_to_string` input arg for converting data to a string. This simplifies the use of `get()` by not having to perform additional processing after the data is fetched. Additionally, some code refactor and fixes were added. Also, the documentation was updated.

New features

- Add input arg to convert output data to a string for the get() method (97)

Changes

- Refactor the api code (91)
- Code refactor for the data loader classes (92)
- Add pipenv to the project (94)
- Add a [gitter room](https://gitter.im/dbcollection/dbcollection) and badge (95)
- Update docs (99)

Fixes

- Solve name collision in SetLoader class (85)

Reached the `v0.2.0` milestone.

With this milestone, the api is more stable and contains more features, as well as improved documentation.

This release brings the `to_memory` feature for moving data fields to memory / disk and lots of bug fixes. Also, the some documentation has been updated with new sections being written. This version bumped to `0.1.11` from `0.1.8` due to some mis-clicks and errors in distributing `mnist`'s data files with `pypi` and `conda`.

New features
- Feature add data allocation to memory (67)
- Add url checking of dataset's source files to CI (68)
- Add lint checking (65)

Changes
- Update docs (71, 75)
- Remove dataset name requirement from the data files' path (78)

Bug fixes
- Fix deploy (64)
- Fix deleting task file (66)
- Fix methods `info_cache()` and `info_datasets()` (76)
- Fix `query()` API method (77)
- Other small bug fixes