Install scripts, bdist_wheel, x86 docker and instructions

[passalis]

This PR adds the following:

• Scripts to install OpenDR toolkit in clean Ubuntu 20.04 systems (even when running from a minimal image, e.g,. docker ones)
• setup.py to correctly install OpenDR package
• Corrected scripts to activate OpenDR venv environment
• Scripts to create bdist wheels for cpu only usage
• Dockerfile for assembling cpu-only OpenDR inference
• Readme listing different installation options
• Update wiki to reflect the changes made in this PR

This PR also add missing __init__.py in toolkit and a __version__ variable according to typical python usage.

[vniclas]

Hi @passalis, the majority of my comments has been addressed.

Concerning 'requires a lot of RAM', I would rephrase the current sentence make sure that you have enough RAM available with a guess of how much will be actually required.

With respect to the pip packages, detectron2 does not actually trigger overwriting the installed versions, i.e., as long as detectron2 still works as expected, this probably does not matter.

Another thing that came to my mind is that it would probably be good to add brief instructions for the user on how to run unittests to verify the successful installation of the toolkit.

[passalis]

I think bcolz was not built correctly in gpu docker image. I am checking this right now.

Actually, it works ok. I was testing a previous image. The image currently uploaded on dockerhub seems to work ok.

[passalis]

@jelledouwe @vniclas thanks for your response!

Concerning 'requires a lot of RAM', I would rephrase the current sentence make sure that you have enough RAM available with a guess of how much will be actually required.

I will run some tests now and measure memory consumption.

Another thing that came to my mind is that it would probably be good to add brief instructions for the user on how to run unittests to verify the successful installation of the toolkit.

I will add some instructions for running the tests to verify the installation.

[passalis]

I added a RAM estimate for installation and updated installation instructions for pip wheel, as well as for running the tests.

[ad-daniel]

I've moved the instructions to build your own docker image to its own section to avoid confusion.
Also, before merging this PR we need to remember to change all mentions of the install_scripts branch

[passalis]

I added some instructions for connecting the local X server into the docker (e.g., for displaying OpenCV windows). Also, I removed all references to install_scripts so some commands/instructions are now broken. As soon as we merge this PR, I will test everything again to make sure that they work as intended.

[ad-daniel]

For it to work, personally I had to run xhost +local:root on my host machine. Otherwise I was getting qt.qpa.xcb: could not connect to display unix:1 qt.qpa.plugin: Could not load the Qt platform plugin "xcb" in "/opendr/venv/lib/python3.8/site-packages/cv2/qt/plugins" even though it was found. when trying to run detr's inference_demo. But could be just me.

Other than that, the PR looks good to me.

@jelledouwe if your points have been addressed could you approve the PR or we won't be able to merge it if you have pending request changes (and also we need 2 approvals)

[passalis]

You are right @ad-daniel I also had to run this command, but I though it was not needed eventually. So I revised the instructions.

[ad-daniel]

Thank you!

Changes made