Add object_type to deepstack.object_detected

Adds object_type to the event deepstack.object_detected. Objects are grouped by the following object_type:

• person: person
• animal: bird, cat, dog, horse, sheep, cow, elephant, bear, zebra, giraffe
• vehicle: bicycle, car, motorcycle, airplane, bus, train, truck
• other: any object that is not in person, animal or vehicle

This will simplify automations, e.g. if you are watching for dog or cat (both need to be in targets), you can now trigger an alert on object_type : animal

Fix small issue

Fixes #173

Adds saved_file to object_detected event

Adds saved_file path information to object_detected event to make including saved images in automations easier. An example is given in the readme

Resolve issue with custom models

Closes #165

Drop unit of measurement "targets"

Drop unit of measurement "targets" and tidies attributes which were way too busy

Bump dependency

Update dependency, no functional changes

Support custom models

Updates readme to reflect open sourcing of Deepstack and adds support for custom models. Note custom models are not supplied, you must create your own or source one from the community.

Example mask detection using a custom model:

adds show_boxes

Adds a new config variable show_boxes. This is a boolean which is True by default, but if configured as False then bounding boxes are not shown on saved images. Useful if you want unadulaterated images saved

Better error handling

As title

Fix naming of saved images

Slight change to the naming of timestamped saved files, the : were causing issues on samba so are replaced with - An example of the new filename is deepstack_object_local_file_1_2020-05-21_04-01-13.jpg
Also a dependency bump

Closes #123