Proposed changes to the calibration documentation

[valenipes]

Hi, this is my first pull request, so I apologize if I made any mistakes in the process. While setting up the integration, I spent a lot of time trying to understand how the MQTT Vacuums Camera was supposed to pass calibration data to the card. To simplify the setup process, I made some changes to the documentation and also made a few minor adjustments.

[sca075]

@valenipes thanks but please point the user to this doccument instead of the valetudo.cloud. I'm not associated in any sense with "valetudo",the camera is a project that live separately from that. In the document it is explained how to get the topic from the attributes of the camera. Anyhow, the setup here is a generic guide. Thanks for your understanding an cooperation.

[valenipes]

Hi @sca075, I understand and I can change it of course.

I've been thinking about the whole doc page and there are some things that I don't understand: the original file claims that for "just room cleaning" it's sufficient identity as calibration source.
First of all, in my experience this isn't true, because identity as calibration source never worked, calibration was wrong. I don't know if it's because of some misconfiguration, but I didn't change much in original valetudo's settings.
Then there's written: "In other cases you have to manually calibrate the map and provide calibration points". I never did manual calibration, but written as this it seems that this is the only alternative, and of course this is not true because one can also use camera as calibration source, and it's way easier (but this isn't written in the original docs).
Then, in the Requirements section, it uses camera as calibration source in the example. To me this is all a bit confusing. Maybe, since I've already to change that little thing, I can also give a better order to the whole doc page based on my experience?
Please let me know.

[sca075]

First of all, in my experience this isn't true, because identity as calibration source never worked, calibration was wrong.

mmmm that is a little strange as per if you use the MQTT Vacuum Camera the calibration_points are in the attributes and if you set camera as source then it should work..

the source is :

map_source:
  camera: camera.valetudo_s5_glossyhardtofindnarwhal_camera
calibration_source:
  camera: true
entity: vacuum.valetudo_glossyhardtofindnarwhal

unless, you don't see the image or an error occurs, anyhow you could have open a discussion or issue and I would have help you ;)
If you are using something different to render the images, well at this point okay but message bottom line. MQTT Vacuum Camera is not associate to Valetudo.cloud (that it is a complete different project), so please reference one or the other but not link them.

[PiotrMachowski]

@sca075 I think that @valenipes meant that calibration_source: { identity: true } never worked.

It should work without any issues, just not with autogenerated rooms.
This documentation was created before great mqtt_vacuum_camera custom integration was created by @sca075, so some parts of it are not necessarily up to date.

[valenipes]

Exactly @PiotrMachowski. At the end I figured out that calibration_source: { identity: true } was the culprit with autogenerated rooms, because when I was trying to get coordinates (following this guide) I received values that were completely different from those of the rooms. But changing to calibration_source: { camera: true } solved all the issues, as @sca075 said. So, in my opinion, for a newcomer (like me) it would be less confusing to directly suggest to autogenerate rooms and then use calibration_source: { camera: true } with mqtt_vacuum_camera. I don't know if it makes sense