Extended Dlib image_processing to save images to directories and bugfixes

[tchellomello]

Description:

During the image_processing.dlib* configuration process I found a little bit complicated to understand if the component was working or not.
So the idea of this PR is to make the process a little more comprehensive by allowing the user some options to save the images captures by Dlib to a directory to determine how accurate the face_recognition is.

This PR also fixes the issue reported at #8493 since dlib behaves a little bit different than microsoft_face_recognition.

This also allows dlib_face_identify to show the name of the face detected instead showing the meta object on the frontend.

• Fixed traceback AttributeError: 'tuple' object has no attribute 'update' (fixes dlib: AttributeError: 'tuple' object has no attribute 'update' #8493)
• Introduced ability to debug picture when a face is or not detected.
• Treats exception when image cannot be parsed
• Modified state_attributes to return the face_name attribute instead object

Related issue (if applicable): fixes #8493

Pull request in home-assistant.github.io with documentation (if applicable): home-assistant/home-assistant.github.io#<home-assistant.github.io PR number goes here> wip

Example entry for configuration.yaml (if applicable):

#configuration.yaml
image_processing: !include image_processing.yaml
automation: !include automations.yaml
android_ip_webcam:
  - host: 192.168.169.121
    scan_interval: 3

#image_processing.yaml
- platform: dlib_face_detect
  keep_faces: True
  keep_no_faces: True
  source:
    - entity_id: camera.ip_webcam
      name: "detect face"

- platform: dlib_face_identify
  keep_known_faces: True
  keep_unknown_faces: True
  source:
    - entity_id: camera.ip_webcam
      name: "Identify mmello"
  faces:
    mmello: /home/hass/.homeassistant/faces/mmello.png
    mmello-fail: /home/hass/.homeassistant/faces/mmello-fail.jpg

#automation.yaml
- alias: mmello detected
  trigger:
    platform: event
    event_type: image_processing.detect_face
    event_data:
      entity_id: image_processing.identify_mmello
      name: mmello
  action:
    service: homeassistant.turn_on
    entity_id: switch.kitchen_light

When enabled, the options above will save the images on the directories {CONFIGURATION_HASS}/:

(ha-py36) ↪ ls -lad dlib_*
drwxrwxr-x. 2 mdemello mdemello 12288 Jul 17 01:23 dlib_faces/
drwxrwxr-x. 2 mdemello mdemello  4096 Jul 17 01:23 dlib_known_faces/
drwxrwxr-x. 2 mdemello mdemello 12288 Jul 17 01:23 dlib_nofaces/
drwxrwxr-x. 2 mdemello mdemello 12288 Jul 17 01:23 dlib_unknown_faces/

(ha-py36) ↪ ls -la dlib_known_faces/ | head
total 1296
drwxrwxr-x. 2 mdemello mdemello  4096 Jul 17 01:23 .
drwxrwxr-x. 9 mdemello mdemello  4096 Jul 17 01:23 ..
-rw-rw-r--. 1 mdemello mdemello 16673 Jul 17 00:36 camera.ip_webcam_Jul-17-2017_00:36:14.jpg
-rw-rw-r--. 1 mdemello mdemello 18964 Jul 17 00:36 camera.ip_webcam_Jul-17-2017_00:36:36.jpg
-rw-rw-r--. 1 mdemello mdemello 18681 Jul 17 00:36 camera.ip_webcam_Jul-17-2017_00:36:47.jpg
-rw-rw-r--. 1 mdemello mdemello 18046 Jul 17 00:36 camera.ip_webcam_Jul-17-2017_00:36:58.jpg
-rw-rw-r--. 1 mdemello mdemello 17850 Jul 17 00:37 camera.ip_webcam_Jul-17-2017_00:37:09.jpg
-rw-rw-r--. 1 mdemello mdemello 16434 Jul 17 00:37 camera.ip_webcam_Jul-17-2017_00:37:42.jpg
-rw-rw-r--. 1 mdemello mdemello 16771 Jul 17 00:38 camera.ip_webcam_Jul-17-2017_00:38:04.jpg

Checklist:

If user exposed functionality or configuration variables are added/changed:

• Documentation added/updated in home-assistant.github.io WIP

If the code does not interact with devices:

• Local tests with tox run successfully. Your PR cannot be merged unless tests pass
• Tests have been added to verify that the new code works.

[pvizeli]

It is not the Task of the image_processing to store image. @fabaff is working on a snapshot servcies, so you can trigger that service with a automation.

I can not see the real fix for that bug. I will make it soon. Need only 1 line code.

[arsaboo]

@pvizeli The problem with the way the image_processing components are currently set up is that there is no feedback to the user about what is going on. I have been using both OpenCV and Dlib and feel like these are black boxes. It would be good to know what was detected so that I can act accordingly. The coordinates in the image_processing entities are of little value to users. Also, saving images/video allow us to use them in notifications, which make these components significantly more useful.

May be there is a better way of handling the image storage, but we definitely need something like this.