spaCy is great as a preprocessor for NLP labeling functions, but there are other libraries that individuals may want to use.

Ideally, we'd like to have wrappers for other packages as well, such as Stanford CoreNLP (https://stanfordnlp.github.io/stanfordnlp/) and NLTK (https://www.nltk.org/). We can pattern match on the [SpacyPreprocessor](https://snorkel.readthedocs.io/en/master/packages/_autos

Thank you for the great effort, you are putting into this project :) There is, however, a feature I miss; rotated bounding boxes. Especially when objects are thin and diagonal, an ordinary bounding box fits poorly. Examples of such cases are shown here: rotated bounding boxes

A way annotation could be

Describe the bug
Project settings -> tags, are not saved unless user hits 'Enter' after typing new tag.

To Reproduce
Steps to reproduce the behavior:

1. In project settings, add new tag.
2. Click 'Save Project'.
3. Exit project settings pane and return.
4. New tag is not saved.

Expected behavior
Tag should be saved when clicking 'Save Project'.

Screenshots
If a

We have a lot of repetitive instances in our data and it a simple Copy/Cut - Paste (Ctrl+C/X and Ctrl+V) function for bounding boxes would help speed up labelling by a lot!

Edit: Also multi-selection of multiple boxes/geometries would be great for a faster workflow

Is your feature request related to a problem? Please describe.

Can't delete annotation types.

Describe the solution you'd like

Allow deletion of annotation types in the web-ui.

Workaround

$ sudo 

This affects the COCO and Pascal VOC exporters. Currently the output class ID is dependent on the order in which they were created in deeplabel. It may be preferable to, in some cases, force this to something else, for example provide a pbtxt file like https://github.com/tensorflow/models/blob/master/research/object_detection/data/mscoco_label_map.pbtxt

The darknet exporter does not have this p

Current Behaviour

• there are no docs with SCSS styles that would help developers to write code in the same style

Expected Behaviour

• we should create docs with code styles we would like to use, providing useful links for everyone to extend their clean code knowledge

Current distribution method involves providing source only via git clone. This package should be pip installable, and ideally have a conda package that can be installed as well.

From @dpordomingo and @mcuadros comments:

• explain better the features provided by our CLU and GUI #29 (comment),
• assume no context/background in the people reading the README.md #29 (comment), -> https://github.com/src-d/code-annotation/p